Datasets:

---

bigcode

/

jupyter-parsed

like 4

Follow

BigCode 1.35k

[ "FTL", "RSA-MD" ]

[ "FTL", "RSA-MD" ]

[ "FTL", "RSA-MD" ]

[ [ [ "# Short instruction to make sure we are working in the right environment\n!conda info", "_____no_output_____" ], [ "#import libraries\nimport os, io\nimport tempfile\nimport time\nimport re\nimport pdf2image\nfrom pdf2image import convert_from_path\nfrom pdf2image import convert_from_bytes\n\nfrom PyPDF2 import PdfFileReader\n\nfrom IPython.display import display, Image\nimport pytesseract\nimport tesserocr\nimport pdftotext\n\nimport pikepdf\nfrom pikepdf import Pdf, Page\nimport pdfminer\nimport pdfplumber\n\nfrom pdfminer.pdfparser import PDFParser\nfrom pdfminer.pdfdocument import PDFDocument\nfrom pdfminer.converter import PDFPageAggregator\nfrom pdfminer.pdfinterp import PDFResourceManager, PDFPageInterpreter\nfrom pdfminer.converter import TextConverter\nfrom pdfminer.layout import LAParams, LTTextBox, LTTextLine\nfrom pdfminer.pdfpage import PDFPage\nfrom pdfminer.pdfpage import PDFTextExtractionNotAllowed\n\n# to import from google drive\nimport pickle\nimport os.path\nfrom googleapiclient import discovery\nfrom googleapiclient.discovery import build\nfrom googleapiclient import errors\nfrom google_auth_oauthlib.flow import InstalledAppFlow\nfrom google.auth.transport.requests import Request\nfrom googleapiclient.http import MediaIoBaseDownload\nfrom oauth2client import client\nfrom oauth2client import tools\nfrom oauth2client.file import Storage\nfrom PIL import Image", "_____no_output_____" ], [ "# If modifying these scopes, delete the file token.pickle.\nSCOPES = ['https://www.googleapis.com/auth/drive']\n\n\"\"\"Shows basic usage of the Drive v3 API.\nPrints the names and ids of the first 10 files the user has access to.\n\"\"\"\ncreds = None\n# The file token.pickle stores the user's access and refresh tokens, and is\n# created automatically when the authorization flow completes for the first\n# time.\nif os.path.exists('token.pickle'):\n with open('token.pickle', 'rb') as token:\n creds = pickle.load(token)\n# If there are no (valid) credentials available, let the user log in.\nif not creds or not creds.valid:\n if creds and creds.expired and creds.refresh_token:\n creds.refresh(Request())\n else:\n flow = InstalledAppFlow.from_client_secrets_file(\n 'client_secret.json', SCOPES)\n creds = flow.run_local_server(port=0)\n # Save the credentials for the next run\n with open('token.pickle', 'wb') as token:\n pickle.dump(creds, token)\n\nservice = build('drive', 'v3', credentials=creds)\n\n# Call the Drive v3 API\nresults = service.files().list(\n pageSize=10, fields=\"nextPageToken, files(id, name)\").execute()\nitems = results.get('files', [])\n\nif not items:\n print('No files found.')\nelse:\n print('Files:')\n for item in items:\n print(u'{0} ({1})'.format(item['name'], item['id']))\n", "Files:\nCopy of Text_extraction_example.ipynb (1ce-FVwyzcNuIXJYNQ1qybKpMSJ6Yn-5V)\nWork_summary_jordi_201122.xlsx (1KCWqsZ3aUAARRX-6fYcV5rTmxRQulsWp)\nWork reccoring (1cAorJ-zPaorVAI7KmwDtC5pd0v7w-PS2)\nOCR.ipynb (1hWljx3wduIk1xJGdkQil2lID7H_Tkvd8)\nNormativaForestal_Chile.pdf (1PJVk7utukeA1Y0riPo_53gAcCCqQ4D3H)\nPrograma_comision_forestal_mex.pdf (10jCcgdxWFoO0AaQXwWk7RsA0RD77n4la)\nCopy of 11255.pdf (1HdCavlCXP-tAlaAmjPOm-i7efxUsiIsw)\nCopy of Nº 004-2020-P-CE-PJ.pdf (1Y5jdAcakQfIP4A5H06c6lQoNBCFk9-me)\nCopy of 3000.pdf (1rwaPoFock58hHlWIot3KEnUTuEl1p6n7)\nDocument de prova per pdf.pdf (1Z5-HSv2Ktox4UGLMM9MtJgJFCwZ5c70Z)\n" ], [ "# 'application/vnd.google-apps.spreadsheet' and parents in '{}'\".format(folder_id)\n# \"mimeType = 'application/vnd.google-apps.pdf' and parents in '{}'\".format(folder_id)\nfolder_id = \"1JU2YWwn88_0hwP5EYuF5FB2LLTYVZKND\"\nfiletype = \"application/pdf\"\nquery = f\"'{folder_id}' in parents\"\nquery = f\"mimeType='{filetype}'\"\nquery = f\"'{folder_id}' in parents and mimeType='{filetype}'\"\npage_token = None\nwhile True:\n response = service.files().list(q = query,\n spaces=\"drive\",\n fields=\"nextPageToken, files(id, name, mimeType)\",\n pageToken=page_token).execute()\n for file in response.get('files', []):\n # Process change\n print('Found file: %s (%s)' % (file.get('name'), file.get('id')))\n page_token = response.get('nextPageToken', None)\n if page_token is None:\n break", "Found file: PDF_995_dummy.pdf (1BewQJdPcBesbYpS0Y-SgsmxlhSBVo363)\nFound file: dummy.pdf (1wls-EM-h4eOeMJTWLrT21DjUGgsvSKGe)\nFound file: Programa_comision_forestal_mex.pdf (10jCcgdxWFoO0AaQXwWk7RsA0RD77n4la)\nFound file: Copy of 11255.pdf (1HdCavlCXP-tAlaAmjPOm-i7efxUsiIsw)\nFound file: Copy of Nº 004-2020-P-CE-PJ.pdf (1Y5jdAcakQfIP4A5H06c6lQoNBCFk9-me)\nFound file: Copy of 3000.pdf (1rwaPoFock58hHlWIot3KEnUTuEl1p6n7)\nFound file: Document de prova per pdf.pdf (1Z5-HSv2Ktox4UGLMM9MtJgJFCwZ5c70Z)\nFound file: Decreto_ejecutivo_57.pdf (1egVmcj6Ssiupf1mX3BEDBUkYBx7r1tfU)\nFound file: Decreto_ejecutivo.pdf (1_qryhaX0h8obAyLsYzf-u_ph0WVTrL1X)\nFound file: ORDENANZA REGULADORA DE TALA Y PODA DE NEJAPA.pdf (1QMPmYxavfGy2FiWc9yOuxRyfyEu66aj0)\n" ], [ "file_id = ['1wls-EM-h4eOeMJTWLrT21DjUGgsvSKGe', '1BewQJdPcBesbYpS0Y-SgsmxlhSBVo363', '10jCcgdxWFoO0AaQXwWk7RsA0RD77n4la', '1egVmcj6Ssiupf1mX3BEDBUkYBx7r1tfU', '1Z5-HSv2Ktox4UGLMM9MtJgJFCwZ5c70Z', '1HdCavlCXP-tAlaAmjPOm-i7efxUsiIsw', '1Y5jdAcakQfIP4A5H06c6lQoNBCFk9-me', '1rwaPoFock58hHlWIot3KEnUTuEl1p6n7']\nfilename = \"Decreto_ejecutivo_57.pdf\"\n\nrequest = service.files().get_media(fileId=file_id[1])\n\nfh = io.BytesIO()\ndownloader = MediaIoBaseDownload(fh, request)\ndone = False\nwhile done is False:\n status, done = downloader.next_chunk()\n print(\"Download %d%%.\" % int(status.progress() * 100))", "Download 100%.\n" ], [ "# pdf = PdfFileReader(fh)\n# for page in range(0, pdf.getNumPages()):\n# pdf_page = pdf.getPage(page) #Retrieve the content of each page\n# pdf_content = pdf_page.extractText()\n# if pdf_content == \"\":\n# print(\"no text\")", "_____no_output_____" ], [ "# pdf_content", "_____no_output_____" ], [ "pdf = Pdf.open('../tests/resources/fourpages.pdf')\nextracted_text = \"\"\nparser = PDFParser(fh)\ndocument = PDFDocument(parser)\nif not document.is_extractable:\n raise PDFTextExtractionNotAllowed\nsio = io.StringIO()\nrsrcmgr = PDFResourceManager()\nlaparams = LAParams()\ndevice = PDFPageAggregator(rsrcmgr, sio, laparams=laparams)\ninterpreter = PDFPageInterpreter(rsrcmgr, device)\n\n\nfor page in PDFPage.create_pages(document):\n interpreter.process_page(page)\ntext = sio.getvalue()\n \n# layout = device.get_result()\n# for lt_obj in layout:\n# if isinstance(lt_obj, LTTextBox) or isinstance(lt_obj, LTTextLine):\n# extracted_text += lt_obj.get_text()\n# else:\n# print(\"no text to extract\")", "_____no_output_____" ], [ "file_id = '1egVmcj6Ssiupf1mX3BEDBUkYBx7r1tfU'\nfilename = \"Decreto_ejecutivo_57.pdf\"\ndef open_file_from_Gdrive(file_id):\n request = service.files().get_media(fileId=file_id)\n\n fh = io.BytesIO()\n downloader = MediaIoBaseDownload(fh, request)\n done = False\n while done is False:\n status, done = downloader.next_chunk()\n print(\"Download %d%%.\" % int(status.progress() * 100))\n scan = convert_from_bytes(fh.getvalue())\n fh.close()\n \ndef check_file_content(bytesIO):\n bytesIO.getValue()", "_____no_output_____" ], [ "extracted_text", "_____no_output_____" ], [ "fh.read()", "_____no_output_____" ], [ " gradient = np.linspace(0, 1, 256)\n gradient = np.vstack((gradient, gradient))\n image_data = cmap(gradient, bytes=True)\n image = Image.fromarray(image_data, 'RGBA')\n\n # ostream = io.FileIO('../cmaps/' + cmap_name + '.png', 'wb')\n # image.save(ostream, format='PNG')\n # ostream.close()\n\n ostream = io.BytesIO()\n image.save(ostream, format='PNG')\n cbar_png_bytes = ostream.getvalue()\n ostream.close()\n\n cbar_png_data = base64.b64encode(cbar_png_bytes)\n cbar_png_bytes = cbar_png_data.decode('unicode_escape')\n\n return cbar_png_bytes ", "_____no_output_____" ], [ "# Mouting drive Google Drive\nfrom google.colab import drive\ndrive.mount('/content/drive/')", "_____no_output_____" ], [ "filename = \"Decreto_ejecutivo_57.pdf\" \n# inp_path = \"/content/drive/MyDrive/WRI-LatinAmerica-Talent/Preprocessing/raw/\"\n# out_path = \"/content/drive/MyDrive/WRI-LatinAmerica-Talent/Preprocessing/OCR/outputs\"\n# filepaths = [os.path.join(inp_path, file) for file in os.listdir(inp_path)]\n# filepaths", "_____no_output_____" ] ], [ [ "# Checking whether the files are scanned images or true pdfs", "_____no_output_____" ] ], [ [ "def is_image(file_path):\n with open(file_path, \"rb\") as f:\n return pdftotext.PDF(f)\n \n", "_____no_output_____" ], [ "print(is_image(filename))", "_____no_output_____" ] ], [ [ "# Converting pdf to image files and improving quality", "_____no_output_____" ] ], [ [ "def get_image1(file_path):\n \"\"\"Get image out of pdf file_path. Splits pdf file into PIL images of each of its pages.\n \"\"\"\n return convert_from_path(file_path, 500)\n\n# Performance tips according to pdf2image: \n# Using an output folder is significantly faster if you are using an SSD. Otherwise i/o usually becomes the bottleneck.\n# Using multiple threads can give you some gains but avoid more than 4 as this will cause i/o bottleneck (even on my NVMe SSD!).", "_____no_output_____" ], [ "pages = get_image1(filepaths[0])\ndisplay(pages[0])", "_____no_output_____" ] ], [ [ "What can we do here to improve image quality? It already seems pretty good!", "_____no_output_____" ], [ "# Evaluating extraction time from each method and saving text to disk", "_____no_output_____" ] ], [ [ "def export_ocr(text, file, extract, out=out_path):\n \"\"\" Export ocr output text using extract method to file at out\n \"\"\"\n filename = f'{os.path.splitext(os.path.basename(file))[0]}_{extract}.txt'\n with open(os.path.join(out, filename), 'w') as the_file:\n the_file.write(text)\n\ndef wrap_pagenum(page_text, page_num):\n \"\"\" Wrap page_text with page_num tag\n \"\"\"\n return f\"<p n={page_num}>\" + page_text + \"</p>\"", "_____no_output_____" ], [ "# pytesseract extraction\nstart_time = time.time()\nfor file in filepaths:\n pages = get_image1(file)\n text = \"\"\n for pageNum, imgBlob in enumerate(pages):\n page_text = pytesseract.image_to_string(imgBlob, lang=\"spa\")\n text += wrap_pagenum(page_text, pageNum)\n export_ocr(text, file, \"pytesseract\") # write extracted text to disk\nprint(\"--- %s seconds ---\" % (time.time() - start_time))", "_____no_output_____" ], [ "# tesserocr extraction\nstart_time = time.time()\nfor file in filepaths:\n pages = get_image1(file)\n text = \"\"\n for pageNum, imgBlob in enumerate(pages):\n page_text = tesserocr.image_to_text(imgBlob, lang=\"spa\")\n text += wrap_pagenum(page_text, pageNum)\n export_ocr(text, file, \"tesserocr\") # write extracted text to disk\nprint(\"--- %s seconds ---\" % (time.time() - start_time))", "_____no_output_____" ], [ "# tesserocr extraction using the PyTessBaseAPI\nstart_time = time.time()\nfor file in filepaths:\n pages = get_image1(file)\n text = \"\"\n with tesserocr.PyTessBaseAPI(lang=\"spa\") as api:\n for pageNum, imgBlob in enumerate(pages):\n api.SetImage(imgBlob)\n page_text = api.GetUTF8Text()\n text += wrap_pagenum(page_text, pageNum)\n export_ocr(text, file, \"tesserocr_pytess\") # write extracted text to disk\nprint(\"--- %s seconds ---\" % (time.time() - start_time))", "_____no_output_____" ] ], [ [ "It seems that the pytesseract package provides the fastest extraction and by looking at the extracted text it doesn't seem to exist any difference in the output of all the tested methods.", "_____no_output_____" ] ], [ [ "# comparison between text extracted by the different methods\nos.listdir(out_path)", "_____no_output_____" ], [ "# TODO: perform a more programatical comparison between extracted texts", "_____no_output_____" ] ], [ [ "# Let's look at the extracted text", "_____no_output_____" ] ], [ [ "with open(os.path.join(out_path, 'Decreto_ejecutivo_57_pytesseract.txt')) as text:\n extracted_text = text.read()\nextracted_text", "_____no_output_____" ], [ "# Replace \\x0c (page break) by \\n\n# Match 1 or more occurrences of \\n if preceeded by one occurrence of \\n OR \n# Match 1 or more occurrences of \\s (whitespace) if preceeded by one occurrence of \\n OR \n# Match one occurrence of \\n if it isn't followed by \\n\nprint(re.sub(\"(?<=\\n)\\n+|(?<=\\n)\\s+|\\n(?!\\n)\", \" \", extracted_text.replace(\"\\x0c\", \"\\n\")))", "_____no_output_____" ] ] ]

[ "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# <img style=\"float: left; padding-right: 10px; width: 45px\" src=\"https://raw.githubusercontent.com/Harvard-IACS/2018-CS109A/master/content/styles/iacs.png\"> CS109A Introduction to Data Science \n\n## Standard Section 3: Multiple Linear Regression and Polynomial Regression \n\n**Harvard University**<br/>\n**Fall 2019**<br/>\n**Instructors**: Pavlos Protopapas, Kevin Rader, and Chris Tanner<br/>\n**Section Leaders**: Marios Mattheakis, Abhimanyu (Abhi) Vasishth, Robbert (Rob) Struyven<br/>\n\n<hr style='height:2px'>", "_____no_output_____" ] ], [ [ "#RUN THIS CELL \nimport requests\nfrom IPython.core.display import HTML\nstyles = requests.get(\"http://raw.githubusercontent.com/Harvard-IACS/2018-CS109A/master/content/styles/cs109.css\").text\nHTML(styles)", "_____no_output_____" ] ], [ [ "For this section, our goal is to get you familiarized with Multiple Linear Regression. We have learned how to model data with kNN Regression and Simple Linear Regression and our goal now is to dive deep into Linear Regression.\n\nSpecifically, we will: \n \n- Load in the titanic dataset from seaborn\n- Learn a few ways to plot **distributions** of variables using seaborn\n- Learn about different **kinds of variables** including continuous, categorical and ordinal\n- Perform single and multiple linear regression\n- Learn about **interaction** terms\n- Understand how to **interpret coefficients** in linear regression\n- Look at **polynomial** regression\n- Understand the **assumptions** being made in a linear regression model\n- (Extra): look at some cool plots to raise your EDA game", "_____no_output_____" ], [ "", "_____no_output_____" ] ], [ [ "# Data and Stats packages\nimport numpy as np\nimport pandas as pd\n\n# Visualization packages\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nsns.set()", "_____no_output_____" ] ], [ [ "# Extending Linear Regression\n\n## Working with the Titanic Dataset from Seaborn\n\nFor our dataset, we'll be using the passenger list from the Titanic, which famously sank in 1912. Let's have a look at the data. Some descriptions of the data are at https://www.kaggle.com/c/titanic/data, and here's [how seaborn preprocessed it](https://github.com/mwaskom/seaborn-data/blob/master/process/titanic.py).\n\nThe task is to build a regression model to **predict the fare**, based on different attributes.\n\nLet's keep a subset of the data, which includes the following variables: \n\n- age\n- sex\n- class\n- embark_town\n- alone\n- **fare** (the response variable)", "_____no_output_____" ] ], [ [ "# Load the dataset from seaborn \ntitanic = sns.load_dataset(\"titanic\")\ntitanic.head()", "_____no_output_____" ], [ "# checking for null values\nchosen_vars = ['age', 'sex', 'class', 'embark_town', 'alone', 'fare']\ntitanic = titanic[chosen_vars]\ntitanic.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 891 entries, 0 to 890\nData columns (total 6 columns):\nage 714 non-null float64\nsex 891 non-null object\nclass 891 non-null category\nembark_town 889 non-null object\nalone 891 non-null bool\nfare 891 non-null float64\ndtypes: bool(1), category(1), float64(2), object(2)\nmemory usage: 29.8+ KB\n" ] ], [ [ "**Exercise**: check the datatypes of each column and display the statistics (min, max, mean and any others) for all the numerical columns of the dataset.", "_____no_output_____" ] ], [ [ "## your code here\n\n", "_____no_output_____" ], [ "# %load 'solutions/sol1.py'\nprint(titanic.dtypes)\ntitanic.describe()", "age float64\nsex object\nclass category\nembark_town object\nalone bool\nfare float64\ndtype: object\n" ] ], [ [ "**Exercise**: drop all the non-null *rows* in the dataset. Is this always a good idea?", "_____no_output_____" ] ], [ [ "## your code here\n", "_____no_output_____" ], [ "# %load 'solutions/sol2.py'\ntitanic = titanic.dropna(axis=0)\ntitanic.info()", "<class 'pandas.core.frame.DataFrame'>\nInt64Index: 712 entries, 0 to 890\nData columns (total 6 columns):\nage 712 non-null float64\nsex 712 non-null object\nclass 712 non-null category\nembark_town 712 non-null object\nalone 712 non-null bool\nfare 712 non-null float64\ndtypes: bool(1), category(1), float64(2), object(2)\nmemory usage: 29.3+ KB\n" ] ], [ [ "Now let us visualize the response variable. A good visualization of the distribution of a variable will enable us to answer three kinds of questions:\n\n- What values are central or typical? (e.g., mean, median, modes)\n- What is the typical spread of values around those central values? (e.g., variance/stdev, skewness)\n- What are unusual or exceptional values (e.g., outliers)", "_____no_output_____" ] ], [ [ "fig, ax = plt.subplots(1, 3, figsize=(24, 6))\nax = ax.ravel()\n\nsns.distplot(titanic['fare'], ax=ax[0])\n# use seaborn to draw distributions\nax[0].set_title('Seaborn distplot')\nax[0].set_ylabel('Normalized frequencies')\n\nsns.violinplot(x='fare', data=titanic, ax=ax[1])\nax[1].set_title('Seaborn violin plot')\nax[1].set_ylabel('Frequencies')\n\nsns.boxplot(x='fare', data=titanic, ax=ax[2])\nax[2].set_title('Seaborn box plot')\nax[2].set_ylabel('Frequencies')\nfig.suptitle('Distribution of count');", "_____no_output_____" ] ], [ [ "How do we interpret these plots?", "_____no_output_____" ], [ "## Train-Test Split", "_____no_output_____" ] ], [ [ "from sklearn.model_selection import train_test_split\n\ntitanic_train, titanic_test = train_test_split(titanic, train_size=0.7, random_state=99)\ntitanic_train = titanic_train.copy()\ntitanic_test = titanic_test.copy()\nprint(titanic_train.shape, titanic_test.shape)", "(498, 6) (214, 6)\n" ] ], [ [ "## Simple one-variable OLS", "_____no_output_____" ], [ "**Exercise**: You've done this before: make a simple model using the OLS package from the statsmodels library predicting **fare** using **age** using the training data. Name your model `model_1` and display the summary", "_____no_output_____" ] ], [ [ "from statsmodels.api import OLS\nimport statsmodels.api as sm", "_____no_output_____" ], [ "# Your code here\n\n", "_____no_output_____" ], [ "# %load 'solutions/sol3.py'\nage_ca = sm.add_constant(titanic_train['age'])\nmodel_1 = OLS(titanic_train['fare'], age_ca).fit()\nmodel_1.summary()", "_____no_output_____" ] ], [ [ "## Dealing with different kinds of variables", "_____no_output_____" ], [ "In general, you should be able to distinguish between three kinds of variables: \n\n1. Continuous variables: such as `fare` or `age`\n2. Categorical variables: such as `sex` or `alone`. There is no inherent ordering between the different values that these variables can take on. These are sometimes called nominal variables. Read more [here](https://stats.idre.ucla.edu/other/mult-pkg/whatstat/what-is-the-difference-between-categorical-ordinal-and-interval-variables/). \n3. Ordinal variables: such as `class` (first > second > third). There is some inherent ordering of the values in the variables, but the values are not continuous either. \n\n*Note*: While there is some inherent ordering in `class`, we will be treating it like a categorical variable.", "_____no_output_____" ] ], [ [ "titanic_orig = titanic_train.copy()", "_____no_output_____" ] ], [ [ "Let us now examine the `sex` column and see the value counts.", "_____no_output_____" ] ], [ [ "titanic_train['sex'].value_counts()", "_____no_output_____" ] ], [ [ "**Exercise**: Create a column `sex_male` that is 1 if the passenger is male, 0 if female. The value counts indicate that these are the two options in this particular dataset. Ensure that the datatype is `int`.", "_____no_output_____" ] ], [ [ "# your code here\n\n", "_____no_output_____" ], [ "# %load 'solutions/sol4.py'\n# functions that help us create a dummy variable\nstratify\ntitanic_train['sex_male'].value_counts()", "_____no_output_____" ] ], [ [ "Do we need a `sex_female` column, or a `sex_others` column? Why or why not?\n\nNow, let us look at `class` in greater detail.", "_____no_output_____" ] ], [ [ "titanic_train['class_Second'] = (titanic_train['class'] == 'Second').astype(int)\ntitanic_train['class_Third'] = 1 * (titanic_train['class'] == 'Third') # just another way to do it", "_____no_output_____" ], [ "titanic_train.info()", "<class 'pandas.core.frame.DataFrame'>\nInt64Index: 498 entries, 278 to 805\nData columns (total 9 columns):\nage 498 non-null float64\nsex 498 non-null object\nclass 498 non-null category\nembark_town 498 non-null object\nalone 498 non-null bool\nfare 498 non-null float64\nsex_male 498 non-null int64\nclass_Second 498 non-null int64\nclass_Third 498 non-null int64\ndtypes: bool(1), category(1), float64(2), int64(3), object(2)\nmemory usage: 32.2+ KB\n" ], [ "# This function automates the above:\ntitanic_train_copy = pd.get_dummies(titanic_train, columns=['sex', 'class'], drop_first=True)\ntitanic_train_copy.head()", "_____no_output_____" ] ], [ [ "## Linear Regression with More Variables", "_____no_output_____" ], [ "**Exercise**: Fit a linear regression including the new sex and class variables. Name this model `model_2`. Don't forget the constant!", "_____no_output_____" ] ], [ [ "# your code here", "_____no_output_____" ], [ "# %load 'solutions/sol5.py'\nmodel_2 = sm.OLS(titanic_train['fare'], \n sm.add_constant(titanic_train[['age', 'sex_male', 'class_Second', 'class_Third']])).fit()\nmodel_2.summary()", "_____no_output_____" ] ], [ [ "### Interpreting These Results", "_____no_output_____" ], [ "1. Which of the predictors do you think are important? Why?\n2. All else equal, what does being male do to the fare?\n\n### Going back to the example from class\n\n\n\n3. What is the interpretation of $\\beta_0$ and $\\beta_1$?", "_____no_output_____" ], [ "## Exploring Interactions", "_____no_output_____" ] ], [ [ "sns.lmplot(x=\"age\", y=\"fare\", hue=\"sex\", data=titanic_train, size=6)", "/anaconda3/envs/109a/lib/python3.7/site-packages/seaborn/regression.py:546: UserWarning: The `size` paramter has been renamed to `height`; please update your code.\n warnings.warn(msg, UserWarning)\n" ] ], [ [ "The slopes seem to be different for male and female. What does that indicate?\n\nLet us now try to add an interaction effect into our model.", "_____no_output_____" ] ], [ [ "# It seemed like gender interacted with age and class. Can we put that in our model?\ntitanic_train['sex_male_X_age'] = titanic_train['age'] * titanic_train['sex_male']\n\nmodel_3 = sm.OLS(\n titanic_train['fare'],\n sm.add_constant(titanic_train[['age', 'sex_male', 'class_Second', 'class_Third', 'sex_male_X_age']])\n).fit()\nmodel_3.summary()", "_____no_output_____" ] ], [ [ "**What happened to the `age` and `male` terms?**", "_____no_output_____" ] ], [ [ "# It seemed like gender interacted with age and class. Can we put that in our model?\ntitanic_train['sex_male_X_class_Second'] = titanic_train['age'] * titanic_train['class_Second']\ntitanic_train['sex_male_X_class_Third'] = titanic_train['age'] * titanic_train['class_Third']\n\nmodel_4 = sm.OLS(\n titanic_train['fare'],\n sm.add_constant(titanic_train[['age', 'sex_male', 'class_Second', 'class_Third', 'sex_male_X_age', \n 'sex_male_X_class_Second', 'sex_male_X_class_Third']])\n).fit()\nmodel_4.summary()", "_____no_output_____" ] ], [ [ "## Polynomial Regression \n\n", "_____no_output_____" ], [ "Perhaps we now believe that the fare also depends on the square of age. How would we include this term in our model?", "_____no_output_____" ] ], [ [ "fig, ax = plt.subplots(figsize=(12,6))\nax.plot(titanic_train['age'], titanic_train['fare'], 'o')\nx = np.linspace(0,80,100)\nax.plot(x, x, '-', label=r'$y=x$')\nax.plot(x, 0.04*x**2, '-', label=r'$y=c x^2$')\nax.set_title('Plotting Age (x) vs Fare (y)')\nax.set_xlabel('Age (x)')\nax.set_ylabel('Fare (y)')\nax.legend();", "_____no_output_____" ] ], [ [ "**Exercise**: Create a model that predicts fare from all the predictors in `model_4` + the square of age. Show the summary of this model. Call it `model_5`. Remember to use the training data, `titanic_train`.", "_____no_output_____" ] ], [ [ "# your code here\n\n", "_____no_output_____" ], [ "# %load 'solutions/sol6.py'\ntitanic_train['age^2'] = titanic_train['age'] **2\nmodel_5 = sm.OLS(\n titanic_train['fare'],\n sm.add_constant(titanic_train[['age', 'sex_male', 'class_Second', 'class_Third', 'sex_male_X_age', \n 'sex_male_X_class_Second', 'sex_male_X_class_Third', 'age^2']])\n).fit()\nmodel_5.summary()", "_____no_output_____" ] ], [ [ "## Looking at All Our Models: Model Selection", "_____no_output_____" ], [ "What has happened to the $R^2$ as we added more features? Does this mean that the model is better? (What if we kept adding more predictors and interaction terms? **In general, how should we choose a model?** We will spend a lot more time on model selection and learn about ways to do so as the course progresses.", "_____no_output_____" ] ], [ [ "models = [model_1, model_2, model_3, model_4, model_5]\nfig, ax = plt.subplots(figsize=(12,6))\nax.plot([model.df_model for model in models], [model.rsquared for model in models], 'x-')\nax.set_xlabel(\"Model degrees of freedom\")\nax.set_title('Model degrees of freedom vs training $R^2$')\nax.set_ylabel(\"$R^2$\");", "_____no_output_____" ] ], [ [ "**What about the test data?**\n\nWe added a lot of columns to our training data and must add the same to our test data in order to calculate $R^2$ scores.", "_____no_output_____" ] ], [ [ "# Added features for model 1\n# Nothing new to be added\n\n# Added features for model 2\ntitanic_test = pd.get_dummies(titanic_test, columns=['sex', 'class'], drop_first=True)\n\n# Added features for model 3\ntitanic_test['sex_male_X_age'] = titanic_test['age'] * titanic_test['sex_male']\n\n# Added features for model 4\ntitanic_test['sex_male_X_class_Second'] = titanic_test['age'] * titanic_test['class_Second']\ntitanic_test['sex_male_X_class_Third'] = titanic_test['age'] * titanic_test['class_Third']\n\n# Added features for model 5\ntitanic_test['age^2'] = titanic_test['age'] **2", "_____no_output_____" ] ], [ [ "**Calculating R^2 scores**", "_____no_output_____" ] ], [ [ "from sklearn.metrics import r2_score\n\nr2_scores = []\ny_preds = []\ny_true = titanic_test['fare']\n\n# model 1\ny_preds.append(model_1.predict(sm.add_constant(titanic_test['age'])))\n\n# model 2\ny_preds.append(model_2.predict(sm.add_constant(titanic_test[['age', 'sex_male', 'class_Second', 'class_Third']])))\n\n# model 3\ny_preds.append(model_3.predict(sm.add_constant(titanic_test[['age', 'sex_male', 'class_Second', 'class_Third', \n 'sex_male_X_age']])))\n\n# model 4\ny_preds.append(model_4.predict(sm.add_constant(titanic_test[['age', 'sex_male', 'class_Second', 'class_Third', \n 'sex_male_X_age', 'sex_male_X_class_Second', \n 'sex_male_X_class_Third']])))\n\n# model 5\ny_preds.append(model_5.predict(sm.add_constant(titanic_test[['age', 'sex_male', 'class_Second', \n 'class_Third', 'sex_male_X_age', \n 'sex_male_X_class_Second', \n 'sex_male_X_class_Third', 'age^2']])))\n\nfor y_pred in y_preds:\n r2_scores.append(r2_score(y_true, y_pred))\n \nmodels = [model_1, model_2, model_3, model_4, model_5]\nfig, ax = plt.subplots(figsize=(12,6))\nax.plot([model.df_model for model in models], r2_scores, 'x-')\nax.set_xlabel(\"Model degrees of freedom\")\nax.set_title('Model degrees of freedom vs test $R^2$')\nax.set_ylabel(\"$R^2$\");", "/anaconda3/envs/109a/lib/python3.7/site-packages/numpy/core/fromnumeric.py:2389: FutureWarning: Method .ptp is deprecated and will be removed in a future version. Use numpy.ptp instead.\n return ptp(axis=axis, out=out, **kwargs)\n" ] ], [ [ "## Regression Assumptions. Should We Even Regress Linearly?", "_____no_output_____" ], [ "", "_____no_output_____" ], [ "**Question**: What are the assumptions of a linear regression model? \n\nWe find that the answer to this question can be found on closer examimation of $\\epsilon$. What is $\\epsilon$? It is assumed that $\\epsilon$ is normally distributed with a mean of 0 and variance $\\sigma^2$. But what does this tell us?\n\n1. Assumption 1: Constant variance of $\\epsilon$ errors. This means that if we plot our **residuals**, which are the differences between the true $Y$ and our predicted $\\hat{Y}$, they should look like they have constant variance and a mean of 0. We will show this in our plots.\n2. Assumption 2: Independence of $\\epsilon$ errors. This again comes from the distribution of $\\epsilon$ that we decide beforehand.\n3. Assumption 3: Linearity. This is an implicit assumption as we claim that Y can be modeled through a linear combination of the predictors. **Important Note:** Even though our predictors, for instance $X_2$, can be created by squaring or cubing another variable, we still use them in a linear equation as shown above, which is why polynomial regression is still a linear model.\n4. Assumption 4: Normality. We assume that the $\\epsilon$ is normally distributed, and we can show this in a histogram of the residuals.\n\n**Exercise**: Calculate the residuals for model 5, our most recent model. Optionally, plot and histogram these residuals and check the assumptions of the model.", "_____no_output_____" ] ], [ [ "# your code here", "_____no_output_____" ], [ "# %load 'solutions/sol7.py'\n# %load 'solutions/sol7.py'\npredictors = sm.add_constant(titanic_train[['age', 'sex_male', 'class_Second', 'class_Third', 'sex_male_X_age', \n 'sex_male_X_class_Second', 'sex_male_X_class_Third', 'age^2']])\ny_hat = model_5.predict(predictors)\nresiduals = titanic_train['fare'] - y_hat\n\n# plotting\nfig, ax = plt.subplots(ncols=2, figsize=(16,5))\nax = ax.ravel()\nax[0].set_title('Plot of Residuals')\nax[0].scatter(y_hat, residuals, alpha=0.2)\nax[0].set_xlabel(r'$\\hat{y}$')\nax[0].set_xlabel('residuals')\n\nax[1].set_title('Histogram of Residuals')\nax[1].hist(residuals, alpha=0.7)\nax[1].set_xlabel('residuals')\nax[1].set_ylabel('frequency');\n\n# Mean of residuals\nprint('Mean of residuals: {}'.format(np.mean(residuals)))", "Mean of residuals: 4.784570776163707e-13\n" ] ], [ [ "**What can you say about the assumptions of the model?**", "_____no_output_____" ], [ "----------------\n### End of Standard Section\n---------------", "_____no_output_____" ], [ "## Extra: Visual exploration of predictors' correlations\n\nThe dataset for this problem contains 10 simulated predictors and a response variable. ", "_____no_output_____" ] ], [ [ "# read in the data \ndata = pd.read_csv('../data/dataset3.txt')\ndata.head()", "_____no_output_____" ], [ "# this effect can be replicated using the scatter_matrix function in pandas plotting\nsns.pairplot(data);", "_____no_output_____" ] ], [ [ "Predictors x1, x2, x3 seem to be perfectly correlated while predictors x4, x5, x6, x7 seem correlated.", "_____no_output_____" ] ], [ [ "data.corr()", "_____no_output_____" ], [ "sns.heatmap(data.corr())", "_____no_output_____" ] ], [ [ "## Extra: A Handy Matplotlib Guide", "_____no_output_____" ], [ "\nsource: http://matplotlib.org/faq/usage_faq.html\n\nSee also [this](http://matplotlib.org/faq/usage_faq.html) matplotlib tutorial.", "_____no_output_____" ], [ "\n\nSee also [this](https://mode.com/blog/violin-plot-examples) violin plot tutorial.", "_____no_output_____" ], [ "---", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown", "markdown", "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import glob\nimport os\nfrom collections import Counter\n\nresults_dir = '/datasets/Our_dataset/results'\n\n# load all the .txt files recursively\nall_annotations = glob.glob(results_dir + '/**/*.txt', recursive=True)\nprint(len(all_annotations), 'files')\n", "9182 files\n" ] ], [ [ "## Count all the words", "_____no_output_____" ] ], [ [ "wordcounter = Counter({})\nwords_per_video = []\nfor ann_idx, ann_file in enumerate(all_annotations):\n file = open(ann_file, \"r\")\n words = file.read().split()\n file.close()\n \n current_wordcounter = Counter(words)\n wordcounter += current_wordcounter\n \n words_per_video.append(len(words))\n", "_____no_output_____" ] ], [ [ "## Some stats", "_____no_output_____" ] ], [ [ "print(\"Number of words:\", len(wordcounter))\n\nprint(\"10 most common words:\")\nprint(wordcounter.most_common(10))\n\nprint(\"Max words in a video:\", max(words_per_video))\nprint(\"Min words in a video:\", min(words_per_video))", "Number of words: 9873\n10 most common words:\n[('que', 3723), ('de', 3050), ('la', 2163), ('y', 1988), ('a', 1865), ('el', 1594), ('no', 1589), ('en', 1459), ('un', 1178), ('es', 1169)]\nMax words in a video: 25\nMin words in a video: 1\n" ], [ "words_per_video_counter = Counter(words_per_video)", "_____no_output_____" ], [ "print(words_per_video_counter)", "Counter({11: 762, 12: 746, 9: 662, 10: 650, 13: 643, 8: 602, 5: 601, 4: 592, 7: 570, 6: 549, 14: 524, 3: 513, 2: 438, 15: 380, 1: 329, 16: 267, 17: 179, 18: 92, 19: 35, 20: 21, 21: 15, 22: 8, 23: 2, 25: 1, 24: 1})\n" ] ] ]

[ "code", "markdown", "code", "markdown", "code" ]

[ [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import matplotlib.pyplot as plt\n\nimport pandas as pd\nimport numpy as np\nimport statistics\nimport math\nfrom sklearn.model_selection import train_test_split\nimport random\nimport sklearn\nfrom sklearn import ensemble\nfrom itertools import chain\nfrom typing import TextIO\nimport re\n\nfrom sklearn.model_selection import GridSearchCV\nfrom sklearn.model_selection import RandomizedSearchCV\nfrom sklearn.model_selection import GroupKFold\nfrom sklearn.model_selection import PredefinedSplit\n\nfrom sklearn import linear_model\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.linear_model import ElasticNet\nfrom sklearn.ensemble import RandomForestClassifier\n#from sklearn import svm\n#from sklearn.svm import SVC\n#from sklearn.svm import LinearSVC\nfrom sklearn.ensemble import GradientBoostingClassifier\n#from sklearn.ensemble import VotingRegressor\nfrom sklearn.metrics import mean_squared_error\nfrom sklearn.metrics import roc_auc_score, average_precision_score, roc_curve, precision_recall_curve, confusion_matrix", "_____no_output_____" ], [ "# data\ngenes = ['RPS14', 'CDC5L', 'POLR2I', 'RPS7', 'XAB2', 'RPS19BP1', 'RPL23A', 'SUPT6H', 'PRPF31', 'U2AF1', 'PSMD7',\n 'Hsp10', 'RPS13', 'PHB', 'RPS9', 'EIF5B', 'RPS6', 'RPS11', 'SUPT5H', 'SNRPD2', 'RPL37', 'RPSA', 'COPS6',\n 'DDX51', 'EIF4A3', 'KARS', 'RPL5', 'RPL32', 'SF3A1', 'RPS3A', 'SF3B3', 'POLR2D', 'RPS15A', 'RPL31', 'PRPF19',\n 'SF3B2', 'RPS4X', 'CSE1L', 'RPL6', 'COPZ1', 'PSMB2', 'RPL7', 'PHB2', 'ARCN1', 'RPA2', 'NUP98', 'RPS3', 'EEF2',\n 'USP39', 'PSMD1', 'NUP93', 'AQR', 'RPL34', 'PSMA1', 'RPS27A']\n\ngenes_filter_1 = ['RPS6', 'PRPF19', 'RPL34', 'Hsp10', 'POLR2I', 'EIF5B', 'RPL31',\n 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H', 'EEF2', 'RPS11',\n 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51', 'RPL7', 'RPS9', 'KARS',\n 'SF3A1', 'RPL32', 'PSMB2', 'RPS7', 'EIF4A3', 'U2AF1', 'PSMA1',\n 'PHB', 'POLR2D', 'RPSA', 'RPL23A', 'NUP93', 'AQR', 'RPA2',\n 'SUPT5H', 'RPL6', 'RPS13', 'SF3B2', 'RPS27A', 'PRPF31', 'COPZ1',\n 'RPS4X', 'PSMD1', 'RPS14', 'NUP98', 'USP39', 'CDC5L', 'RPL5',\n 'PHB2', 'RPS15A', 'RPS3', 'ARCN1', 'COPS6']\n\ngene_split_index = {}\nfor i in range(len(genes_filter_1)):\n gene = genes_filter_1[i]\n gene_split_index[gene]= math.floor(i/6)\n\nbase_positions = {\n 'A': 0,\n 'T': 1,\n 'C': 2,\n 'G': 3,\n 0: 'A',\n 1: 'T',\n 2: 'C',\n 3: 'G',\n}\n", "_____no_output_____" ], [ "def create_gene_splits_filter1_kfold_noval(gene_strings, values_to_split: list, kfold, split):\n # use number [0, 1, 2, 3, 4,...] as index\n genes_filter_1 = ['RPS6', 'PRPF19', 'RPL34', 'Hsp10', 'POLR2I', 'EIF5B', 'RPL31',\n 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H', 'EEF2', 'RPS11',\n 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51', 'RPL7', 'RPS9', 'KARS',\n 'SF3A1', 'RPL32', 'PSMB2', 'RPS7', 'EIF4A3', 'U2AF1', 'PSMA1',\n 'PHB', 'POLR2D', 'RPSA', 'RPL23A', 'NUP93', 'AQR', 'RPA2',\n 'SUPT5H', 'RPL6', 'RPS13', 'SF3B2', 'RPS27A', 'PRPF31', 'COPZ1',\n 'RPS4X', 'PSMD1', 'RPS14', 'NUP98', 'USP39', 'CDC5L', 'RPL5',\n 'PHB2', 'RPS15A', 'RPS3', 'ARCN1', 'COPS6']\n assert split >= 0 and split < kfold\n if kfold == 9:\n #val_genes = genes_filter_1[split * 6: (split + 1) * 6]\n if split != 8:\n test_genes = genes_filter_1[((split + 1) * 6): (split + 2) * 6]\n else:\n test_genes = genes_filter_1[0:6]\n #print('val:', val_genes)\n print('test:', test_genes)\n\n #val_ids = list(chain(*[np.where(gene_strings == g)[0] for g in val_genes]))\n test_ids = list(chain(*[np.where(gene_strings == g)[0] for g in test_genes]))\n train_ids = list((set(range(len(gene_strings))) - set(test_ids)))\n\n train = [[arr[i] for i in train_ids] for arr in values_to_split]\n #val = [[arr[i] for i in val_ids] for arr in values_to_split]\n test = [[arr[i] for i in test_ids] for arr in values_to_split]\n\n return train, test\n\ndef normalize(a: np.ndarray):\n \"\"\"\n :param a: numpy array of size N x D, where N is number of examples, D is number of features\n :return: a, normalized so that all feature columns are now between 0 and 1\n \"\"\"\n a_normed, norms = sklearn.preprocessing.normalize(a, norm='max', axis=0, return_norm=True)\n print(\"Norms:\", norms)\n return a_normed\n\ndef one_hot_encode_sequence(seq, pad_to_len=-1):\n output_len = len(seq)\n if pad_to_len > 0:\n assert pad_to_len >= output_len\n output_len = pad_to_len\n\n encoded_seq = np.zeros((output_len, 4), dtype=np.float32)\n for i, base in enumerate(seq):\n encoded_seq[i][base_positions[base]] = 1\n return encoded_seq", "_____no_output_____" ], [ "dataset_filtered_csv_path = '../../../data/integrated_guide_feature_filtered_f24_mismatch3_all_flanks.csv'\n\n#dataset\ndataframe = pd.read_csv(dataset_filtered_csv_path)\ndataframe = dataframe[dataframe['gene'].isin(genes_filter_1)] #filter out 1 gene\n\nnum_examples = len(dataframe['gene'].values)\nencoded_guides = [one_hot_encode_sequence(guide).flatten() for guide in dataframe['guide'].values]\n\n# guide seq only classification\n\nclasses = dataframe['binary_relative_ratio_075f'].values\n\noutputs = classes.astype(np.float32)\n \nall_cols = [encoded_guides, outputs]\n\n# group label to split\ngroups = dataframe['gene'].values\n\n# predefined split index\nfor g in gene_split_index.keys():\n dataframe.loc[dataframe['gene']== g,'predefined split index']= gene_split_index[g]\nps = PredefinedSplit(dataframe['predefined split index'].values)\nprint(ps.get_n_splits())", "9\n" ], [ "len(all_cols[1])", "_____no_output_____" ] ], [ [ "## hp tuning", "_____no_output_____" ] ], [ [ "# LogisticRegression, L1\nlogreg = LogisticRegression(penalty='l1',solver='saga',random_state=0,max_iter=10000)\ngrid = {'C': np.logspace(-5, 5, 11)}\n\n#predefined splits\n#gs = GridSearchCV(logreg, grid, cv=ps.split(),scoring='accuracy')\ngs = GridSearchCV(logreg, grid, cv=ps.split(),scoring=['roc_auc','average_precision'],refit='roc_auc')\ngs.fit(all_cols[0], all_cols[1])\nprint(gs.best_params_)\nprint(gs.best_score_) #best cv score\ndf_gridsearch = pd.DataFrame(gs.cv_results_)\n\ndf_gridsearch.to_csv('model_hp_results/guideonly_gene20_075f_classi_LogisticRegression_L1_hp.csv')", "{'C': 0.1}\n0.7214640628190946\n" ], [ "# LogisticRegression, L2\nlogreg = LogisticRegression(penalty='l2',solver='saga',random_state=0,max_iter=10000)\ngrid = {'C': np.logspace(-5, 5, 11)}\n\n#predefined splits\ngs = GridSearchCV(logreg, grid, cv=ps.split(),scoring=['roc_auc','average_precision'],refit='roc_auc')\ngs.fit(all_cols[0], all_cols[1])\nprint(gs.best_params_)\nprint(gs.best_score_) #best cv score\ndf_gridsearch = pd.DataFrame(gs.cv_results_)\n\ndf_gridsearch.to_csv('model_hp_results/guideonly_gene20_075f_classi_LogisticRegression_L2_hp.csv')", "{'C': 0.01}\n0.7214300947964725\n" ], [ "# LogisticRegression, elasticnet\nlogreg = LogisticRegression(penalty='elasticnet',solver='saga',random_state=0,max_iter=10000)\ngrid = {'C': np.logspace(-4, 4, 9),'l1_ratio':np.linspace(0.1, 1, num=10)}\n\ngs = GridSearchCV(logreg, grid, cv=ps.split(),scoring=['roc_auc','average_precision'],refit='roc_auc')\ngs.fit(all_cols[0], all_cols[1])\nprint(gs.best_params_)\nprint(gs.best_score_) #best cv score\ndf_gridsearch = pd.DataFrame(gs.cv_results_)\n\ndf_gridsearch.to_csv('model_hp_results/guideonly_gene20_075f_classi_LogisticRegression_elasticnet_hp.csv')", "{'C': 0.01, 'l1_ratio': 0.2}\n0.7214917118945797\n" ], [ "# https://www.programcreek.com/python/example/91158/sklearn.model_selection.GroupKFold\n#random forest\nclf = RandomForestClassifier(random_state=0)\ngrid = {'n_estimators':[100,200,400,800,1000,1200,1500],'max_features':['auto','sqrt','log2']}\ngs = GridSearchCV(clf, grid, cv=GroupKFold(n_splits=5))\ngs.fit(all_cols[0], all_cols[1], groups=groups)\n", "_____no_output_____" ], [ "#GradientBoostingClassifier\ngb = ensemble.GradientBoostingClassifier(random_state=0)\ngrid = {'learning_rate':np.logspace(-2, 0, 3),'n_estimators':[100,200,400,800,1000,1200,1500],'max_depth':[2,3,4,8],'max_features':['auto','sqrt','log2']}\ngs = GridSearchCV(gb, grid, cv=GroupKFold(n_splits=5))\ngs.fit(all_cols[0], all_cols[1], groups=groups)\nprint(gs.best_score_) #best cv score\nprint(gs.best_params_)\ndf_gridsearch = pd.DataFrame(gs.cv_results_)\n\ndf_gridsearch.to_csv('linearmodel_hp_results/classi_gb_hp.csv')", "_____no_output_____" ] ], [ [ "## Test models", "_____no_output_____" ] ], [ [ "def classification_analysis(model_name, split, y_pred,y_true):\n test_df = pd.DataFrame(list(zip(list(y_pred), list(y_true))),\n columns =['predicted_value', 'true_binary_label'])\n \n thres_list = [0.8, 0.9,0.95]\n tp_thres = []\n #print('thres_stats')\n for thres in thres_list:\n df_pre_good = test_df[test_df['predicted_value']>thres]\n true_good_label = df_pre_good['true_binary_label'].values\n num_real_gg = np.count_nonzero(true_good_label)\n if len(true_good_label)>0:\n gg_ratio = num_real_gg/len(true_good_label)\n tp_thres.append(gg_ratio)\n #print('true good guide percent '+str(gg_ratio))\n else:\n tp_thres.append('na')\n \n outputs = np.array(y_pred)\n labels = np.array(y_true)\n #plt.clf()\n #fig.suptitle('AUC and PRC')\n score = roc_auc_score(labels, outputs)\n fpr, tpr, _ = roc_curve(labels, outputs)\n #print('AUROC '+str(score))\n average_precision = average_precision_score(labels, outputs)\n precision, recall, thres_prc = precision_recall_curve(labels, outputs)\n #print('AUPRC '+str(average_precision))\n #plt.savefig(fname='results/linear_models/'+str(model_name)+'precision-recall_'+str(split)+'.png',dpi=600,bbox_inches='tight')\n return score,average_precision,tp_thres\n ", "_____no_output_____" ], [ "#LogisticRegression, little regularization\nlogreg = LogisticRegression(penalty='l1',solver='saga',random_state=0,max_iter=10000,C=100000000)\nauroc_l = []\nauprc_l = []\ntp_80 = []\ntp_90 = []\nfor s in range(9):\n #tr, val, te = create_gene_splits_kfold(dataframe['gene'].values, all_cols, 11, s)\n tr, te = create_gene_splits_filter1_kfold_noval(dataframe['gene'].values, all_cols, 9, s)\n # training input and output\n d_input = tr[0]\n d_output = tr[1]\n logreg.fit(d_input, d_output) #fit models\n #test set\n xt = te[0] \n #pred = logreg.predict(xt)\n pred = logreg.predict_proba(xt)\n pred = pred[:,1]\n auroc,auprc,tp_thres = classification_analysis('LogisticRegression-L1', s,pred,te[1])\n auroc_l.append(auroc)\n auprc_l.append(auprc)\n if tp_thres[0]!= 'na':\n tp_80.append(tp_thres[0])\n if tp_thres[1]!= 'na':\n tp_90.append(tp_thres[1])\n \nauroc_mean = statistics.mean(auroc_l)\nauroc_sd = statistics.stdev(auroc_l)\nprint('auroc_mean: '+str(auroc_mean))\nprint('auroc_sd: '+str(auroc_sd))\nauprc_mean = statistics.mean(auprc_l)\nauprc_sd = statistics.stdev(auprc_l)\nprint('auprc_mean: '+str(auprc_mean))\nprint('auprc_sd: '+str(auprc_sd))\n \ntp_80_mean = statistics.mean(tp_80)\ntp_80_sd = statistics.stdev(tp_80)\nprint('tp_80_mean: '+str(tp_80_mean))\nprint('tp_80_sd: '+str(tp_80_sd))\ntp_90_mean = statistics.mean(tp_90)\ntp_90_sd = statistics.stdev(tp_90)\nprint('tp_90_mean: '+str(tp_90_mean))\nprint('tp_90_sd: '+str(tp_90_sd))", "test: ['RPL31', 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H']\ntest: ['EEF2', 'RPS11', 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51']\ntest: ['RPL7', 'RPS9', 'KARS', 'SF3A1', 'RPL32', 'PSMB2']\ntest: ['RPS7', 'EIF4A3', 'U2AF1', 'PSMA1', 'PHB', 'POLR2D']\ntest: ['RPSA', 'RPL23A', 'NUP93', 'AQR', 'RPA2', 'SUPT5H']\ntest: ['RPL6', 'RPS13', 'SF3B2', 'RPS27A', 'PRPF31', 'COPZ1']\ntest: ['RPS4X', 'PSMD1', 'RPS14', 'NUP98', 'USP39', 'CDC5L']\ntest: ['RPL5', 'PHB2', 'RPS15A', 'RPS3', 'ARCN1', 'COPS6']\ntest: ['RPS6', 'PRPF19', 'RPL34', 'Hsp10', 'POLR2I', 'EIF5B']\nauroc_mean: 0.7213890843218597\nauroc_sd: 0.01830150463668443\nauprc_mean: 0.3253363736827444\nauprc_sd: 0.02084537819584787\ntp_80_mean: 0.3\ntp_80_sd: 0.4760952285695233\n" ], [ "# LogisticRegression, L1\nlogreg = LogisticRegression(penalty='l1',solver='saga',random_state=0,max_iter=10000,C=0.1)\nauroc_l = []\nauprc_l = []\ntp_80 = []\ntp_90 = []\nfor s in range(9):\n #tr, val, te = create_gene_splits_kfold(dataframe['gene'].values, all_cols, 11, s)\n tr, te = create_gene_splits_filter1_kfold_noval(dataframe['gene'].values, all_cols, 9, s)\n # training input and output\n d_input = tr[0]\n d_output = tr[1]\n logreg.fit(d_input, d_output) #fit models\n #test set\n xt = te[0] \n #pred = logreg.predict(xt)\n pred = logreg.predict_proba(xt)\n pred = pred[:,1]\n auroc,auprc,tp_thres = classification_analysis('LogisticRegression-L1', s,pred,te[1])\n auroc_l.append(auroc)\n auprc_l.append(auprc)\n if tp_thres[0]!= 'na':\n tp_80.append(tp_thres[0])\n if tp_thres[1]!= 'na':\n tp_90.append(tp_thres[1])\n \nauroc_mean = statistics.mean(auroc_l)\nauroc_sd = statistics.stdev(auroc_l)\nprint('auroc_mean: '+str(auroc_mean))\nprint('auroc_sd: '+str(auroc_sd))\nauprc_mean = statistics.mean(auprc_l)\nauprc_sd = statistics.stdev(auprc_l)\nprint('auprc_mean: '+str(auprc_mean))\nprint('auprc_sd: '+str(auprc_sd))\n \ntp_80_mean = statistics.mean(tp_80)\ntp_80_sd = statistics.stdev(tp_80)\nprint('tp_80_mean: '+str(tp_80_mean))\nprint('tp_80_sd: '+str(tp_80_sd))\ntp_90_mean = statistics.mean(tp_90)\ntp_90_sd = statistics.stdev(tp_90)\nprint('tp_90_mean: '+str(tp_90_mean))\nprint('tp_90_sd: '+str(tp_90_sd))", "test: ['RPL31', 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H']\ntest: ['EEF2', 'RPS11', 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51']\ntest: ['RPL7', 'RPS9', 'KARS', 'SF3A1', 'RPL32', 'PSMB2']\ntest: ['RPS7', 'EIF4A3', 'U2AF1', 'PSMA1', 'PHB', 'POLR2D']\ntest: ['RPSA', 'RPL23A', 'NUP93', 'AQR', 'RPA2', 'SUPT5H']\ntest: ['RPL6', 'RPS13', 'SF3B2', 'RPS27A', 'PRPF31', 'COPZ1']\ntest: ['RPS4X', 'PSMD1', 'RPS14', 'NUP98', 'USP39', 'CDC5L']\ntest: ['RPL5', 'PHB2', 'RPS15A', 'RPS3', 'ARCN1', 'COPS6']\ntest: ['RPS6', 'PRPF19', 'RPL34', 'Hsp10', 'POLR2I', 'EIF5B']\nauroc_mean: 0.7214640628190947\nauroc_sd: 0.018183248177440187\nauprc_mean: 0.32520716564837393\nauprc_sd: 0.020724462125722667\ntp_80_mean: 0.3125\ntp_80_sd: 0.4732423621500228\n" ], [ "# LogisticRegression, L2\nlogreg = LogisticRegression(penalty='l2',solver='saga',random_state=0,max_iter=10000,C=0.01)\nauroc_l = []\nauprc_l = []\ntp_80 = []\ntp_90 = []\nfor s in range(9):\n #tr, val, te = create_gene_splits_kfold(dataframe['gene'].values, all_cols, 11, s)\n tr, te = create_gene_splits_filter1_kfold_noval(dataframe['gene'].values, all_cols, 9, s)\n # training input and output\n d_input = tr[0]\n d_output = tr[1]\n logreg.fit(d_input, d_output) #fit models\n #test set\n xt = te[0] \n #pred = logreg.predict(xt)\n pred = logreg.predict_proba(xt)\n pred = pred[:,1]\n auroc,auprc,tp_thres = classification_analysis('LogisticRegression-L2', s,pred,te[1])\n auroc_l.append(auroc)\n auprc_l.append(auprc)\n if tp_thres[0]!= 'na':\n tp_80.append(tp_thres[0])\n if tp_thres[1]!= 'na':\n tp_90.append(tp_thres[1])\n \nauroc_mean = statistics.mean(auroc_l)\nauroc_sd = statistics.stdev(auroc_l)\nprint('auroc_mean: '+str(auroc_mean))\nprint('auroc_sd: '+str(auroc_sd))\nauprc_mean = statistics.mean(auprc_l)\nauprc_sd = statistics.stdev(auprc_l)\nprint('auprc_mean: '+str(auprc_mean))\nprint('auprc_sd: '+str(auprc_sd))\n \ntp_80_mean = statistics.mean(tp_80)\ntp_80_sd = statistics.stdev(tp_80)\nprint('tp_80_mean: '+str(tp_80_mean))\nprint('tp_80_sd: '+str(tp_80_sd))\ntp_90_mean = statistics.mean(tp_90)\ntp_90_sd = statistics.stdev(tp_90)\nprint('tp_90_mean: '+str(tp_90_mean))\nprint('tp_90_sd: '+str(tp_90_sd))", "test: ['RPL31', 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H']\ntest: ['EEF2', 'RPS11', 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51']\ntest: ['RPL7', 'RPS9', 'KARS', 'SF3A1', 'RPL32', 'PSMB2']\ntest: ['RPS7', 'EIF4A3', 'U2AF1', 'PSMA1', 'PHB', 'POLR2D']\ntest: ['RPSA', 'RPL23A', 'NUP93', 'AQR', 'RPA2', 'SUPT5H']\ntest: ['RPL6', 'RPS13', 'SF3B2', 'RPS27A', 'PRPF31', 'COPZ1']\ntest: ['RPS4X', 'PSMD1', 'RPS14', 'NUP98', 'USP39', 'CDC5L']\ntest: ['RPL5', 'PHB2', 'RPS15A', 'RPS3', 'ARCN1', 'COPS6']\ntest: ['RPS6', 'PRPF19', 'RPL34', 'Hsp10', 'POLR2I', 'EIF5B']\nauroc_mean: 0.7214300947964725\nauroc_sd: 0.0182392049171858\nauprc_mean: 0.3252897453354815\nauprc_sd: 0.02076884919984052\ntp_80_mean: 0.3333333333333333\ntp_80_sd: 0.5773502691896258\n" ], [ "# LogisticRegression, elasticnet\nlogreg = LogisticRegression(penalty='elasticnet',solver='saga',random_state=0,max_iter=10000,l1_ratio=0.50,C=0.1)\nauroc_l = []\nauprc_l = []\ntp_80 = []\ntp_90 = []\nfor s in range(9):\n #tr, val, te = create_gene_splits_kfold(dataframe['gene'].values, all_cols, 11, s)\n tr, te = create_gene_splits_filter1_kfold_noval(dataframe['gene'].values, all_cols, 9, s)\n # training input and output\n d_input = tr[0]\n d_output = tr[1]\n logreg.fit(d_input, d_output) #fit models\n #test set\n xt = te[0] \n #pred = logreg.predict(xt)\n pred = logreg.predict_proba(xt)\n pred = pred[:,1]\n auroc,auprc,tp_thres = classification_analysis('LogisticRegression-elasticnet', s,pred,te[1])\n auroc_l.append(auroc)\n auprc_l.append(auprc)\n if tp_thres[0]!= 'na':\n tp_80.append(tp_thres[0])\n if tp_thres[1]!= 'na':\n tp_90.append(tp_thres[1])\n \nauroc_mean = statistics.mean(auroc_l)\nauroc_sd = statistics.stdev(auroc_l)\nprint('auroc_mean: '+str(auroc_mean))\nprint('auroc_sd: '+str(auroc_sd))\nauprc_mean = statistics.mean(auprc_l)\nauprc_sd = statistics.stdev(auprc_l)\nprint('auprc_mean: '+str(auprc_mean))\nprint('auprc_sd: '+str(auprc_sd))\n \ntp_80_mean = statistics.mean(tp_80)\ntp_80_sd = statistics.stdev(tp_80)\nprint('tp_80_mean: '+str(tp_80_mean))\nprint('tp_80_sd: '+str(tp_80_sd))\ntp_90_mean = statistics.mean(tp_90)\ntp_90_sd = statistics.stdev(tp_90)\nprint('tp_90_mean: '+str(tp_90_mean))\nprint('tp_90_sd: '+str(tp_90_sd))", "test: ['RPL31', 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H']\ntest: ['EEF2', 'RPS11', 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51']\ntest: ['RPL7', 'RPS9', 'KARS', 'SF3A1', 'RPL32', 'PSMB2']\ntest: ['RPS7', 'EIF4A3', 'U2AF1', 'PSMA1', 'PHB', 'POLR2D']\ntest: ['RPSA', 'RPL23A', 'NUP93', 'AQR', 'RPA2', 'SUPT5H']\ntest: ['RPL6', 'RPS13', 'SF3B2', 'RPS27A', 'PRPF31', 'COPZ1']\ntest: ['RPS4X', 'PSMD1', 'RPS14', 'NUP98', 'USP39', 'CDC5L']\ntest: ['RPL5', 'PHB2', 'RPS15A', 'RPS3', 'ARCN1', 'COPS6']\ntest: ['RPS6', 'PRPF19', 'RPL34', 'Hsp10', 'POLR2I', 'EIF5B']\nauroc_mean: 0.721435369649852\nauroc_sd: 0.018237601384323694\nauprc_mean: 0.3252827677556211\nauprc_sd: 0.02078180094189064\ntp_80_mean: 0.3125\ntp_80_sd: 0.4732423621500228\n" ], [ "#SVM, linear\n\n\nclf = svm.SVC(kernel='linear',probability=True,random_state=0,C=0.001)\n\n#clf = LinearSVC(dual= False, random_state=0, max_iter=10000,C=1,penalty='l2')\n\nauroc_l = []\nauprc_l = []\ntp_80 = []\ntp_90 = []\nfor s in range(9):\n #tr, val, te = create_gene_splits_kfold(dataframe['gene'].values, all_cols, 11, s)\n tr, te = create_gene_splits_filter1_kfold_noval(dataframe['gene'].values, all_cols, 9, s)\n # training input and output\n d_input = tr[0]\n d_output = tr[1]\n clf.fit(d_input, d_output) #fit models\n #test set\n xt = te[0] \n pred = clf.predict_proba(xt)\n pred = pred[:,1]\n #pred = clf.predict(xt)\n auroc,auprc,tp_thres = classification_analysis('svm', s,pred,te[1])\n auroc_l.append(auroc)\n auprc_l.append(auprc)\n if tp_thres[0]!= 'na':\n tp_80.append(tp_thres[0])\n if tp_thres[1]!= 'na':\n tp_90.append(tp_thres[1])\n \nauroc_mean = statistics.mean(auroc_l)\nauroc_sd = statistics.stdev(auroc_l)\nprint('auroc_mean: '+str(auroc_mean))\nprint('auroc_sd: '+str(auroc_sd))\nauprc_mean = statistics.mean(auprc_l)\nauprc_sd = statistics.stdev(auprc_l)\nprint('auprc_mean: '+str(auprc_mean))\nprint('auprc_sd: '+str(auprc_sd)) \ntp_80_mean = statistics.mean(tp_80)\ntp_80_sd = statistics.stdev(tp_80)\n#print('tp_80_mean: '+str(tp_80_mean))\n#print('tp_80_sd: '+str(tp_80_sd))\ntp_90_mean = statistics.mean(tp_90)\ntp_90_sd = statistics.stdev(tp_90)\n#print('tp_90_mean: '+str(tp_90_mean))\n#print('tp_90_sd: '+str(tp_90_sd))", "test: ['RPL31', 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H']\ntest: ['EEF2', 'RPS11', 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51']\n" ], [ "# random forest\n#clf = RandomForestClassifier(n_estimators=32,min_samples_split=2, min_samples_leaf=2, max_features='auto',random_state=0)\nclf = RandomForestClassifier(n_estimators=1500,max_features='auto',random_state=0)\nauroc_l = []\nauprc_l = []\ntp_80 = []\ntp_90 = []\nfor s in range(9):\n #tr, val, te = create_gene_splits_kfold(dataframe['gene'].values, all_cols, 11, s)\n #tr, val, te = create_gene_splits_filter1_kfold(dataframe['gene'].values, all_cols, 9, args.split)\n tr, te = create_gene_splits_filter1_kfold_noval(dataframe['gene'].values, all_cols, 9, s)\n # training input and output\n d_input = tr[0]\n d_output = tr[1]\n clf.fit(d_input, d_output) #fit models\n #test set\n xt = te[0] \n #pred = logreg.predict(xt)\n pred = clf.predict_proba(xt)\n pred = pred[:,1]\n auroc,auprc,tp_thres = classification_analysis('random forest', s,pred,te[1])\n auroc_l.append(auroc)\n auprc_l.append(auprc)\n if tp_thres[0]!= 'na':\n tp_80.append(tp_thres[0])\n if tp_thres[1]!= 'na':\n tp_90.append(tp_thres[1])\n \nauroc_mean = statistics.mean(auroc_l)\nauroc_sd = statistics.stdev(auroc_l)\nprint('auroc_mean: '+str(auroc_mean))\nprint('auroc_sd: '+str(auroc_sd))\nauprc_mean = statistics.mean(auprc_l)\nauprc_sd = statistics.stdev(auprc_l)\nprint('auprc_mean: '+str(auprc_mean))\nprint('auprc_sd: '+str(auprc_sd))\n \ntp_80_mean = statistics.mean(tp_80)\ntp_80_sd = statistics.stdev(tp_80)\nprint('tp_80_mean: '+str(tp_80_mean))\nprint('tp_80_sd: '+str(tp_80_sd))\ntp_90_mean = statistics.mean(tp_90)\ntp_90_sd = statistics.stdev(tp_90)\nprint('tp_90_mean: '+str(tp_90_mean))\nprint('tp_90_sd: '+str(tp_90_sd))", "test: ['RPL31', 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H']\ntest: ['EEF2', 'RPS11', 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51']\ntest: ['RPL7', 'RPS9', 'KARS', 'SF3A1', 'RPL32', 'PSMB2']\ntest: ['RPS7', 'EIF4A3', 'U2AF1', 'PSMA1', 'PHB', 'POLR2D']\ntest: ['RPSA', 'RPL23A', 'NUP93', 'AQR', 'RPA2', 'SUPT5H']\ntest: ['RPL6', 'RPS13', 'SF3B2', 'RPS27A', 'PRPF31', 'COPZ1']\ntest: ['RPS4X', 'PSMD1', 'RPS14', 'NUP98', 'USP39', 'CDC5L']\ntest: ['RPL5', 'PHB2', 'RPS15A', 'RPS3', 'ARCN1', 'COPS6']\ntest: ['RPS6', 'PRPF19', 'RPL34', 'Hsp10', 'POLR2I', 'EIF5B']\nauroc_mean: 0.788279464458613\nauroc_sd: 0.01428358067005656\nauprc_mean: 0.4395069525801463\nauprc_sd: 0.023567763312645224\n" ], [ "#GradientBoostingClassifier\nclf = ensemble.GradientBoostingClassifier(random_state=0,max_depth=4,\n max_features='auto', n_estimators=1500)\n\nauroc_l = []\nauprc_l = []\ntp_80 = []\ntp_90 = []\n#for s in range(11):\nfor s in range(9):\n #tr, val, te = create_gene_splits_kfold(dataframe['gene'].values, all_cols, 11, s)\n #tr, val, te = create_gene_splits_filter1_kfold(dataframe['gene'].values, all_cols, 9, args.split)\n tr, te = create_gene_splits_filter1_kfold_noval(dataframe['gene'].values, all_cols, 9, s)\n # training input and output\n d_input = tr[0]\n d_output = tr[1]\n clf.fit(d_input, d_output) #fit models\n #test set\n xt = te[0] \n pred = clf.predict_proba(xt)\n pred = pred[:,1]\n auroc,auprc,tp_thres = classification_analysis('GradientBoostingClassifier_hpnew', s,pred,te[1])\n auroc_l.append(auroc)\n auprc_l.append(auprc)\n if tp_thres[0]!= 'na':\n tp_80.append(tp_thres[0])\n if tp_thres[1]!= 'na':\n tp_90.append(tp_thres[1])\n \nauroc_mean = statistics.mean(auroc_l)\nauroc_sd = statistics.stdev(auroc_l)\nprint('auroc_mean: '+str(auroc_mean))\nprint('auroc_sd: '+str(auroc_sd))\nauprc_mean = statistics.mean(auprc_l)\nauprc_sd = statistics.stdev(auprc_l)\nprint('auprc_mean: '+str(auprc_mean))\nprint('auprc_sd: '+str(auprc_sd))\n \ntp_80_mean = statistics.mean(tp_80)\ntp_80_sd = statistics.stdev(tp_80)\nprint('tp_80_mean: '+str(tp_80_mean))\nprint('tp_80_sd: '+str(tp_80_sd))\ntp_90_mean = statistics.mean(tp_90)\ntp_90_sd = statistics.stdev(tp_90)\nprint('tp_90_mean: '+str(tp_90_mean))\nprint('tp_90_sd: '+str(tp_90_sd))", "test: ['RPL31', 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H']\ntest: ['EEF2', 'RPS11', 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51']\ntest: ['RPL7', 'RPS9', 'KARS', 'SF3A1', 'RPL32', 'PSMB2']\ntest: ['RPS7', 'EIF4A3', 'U2AF1', 'PSMA1', 'PHB', 'POLR2D']\ntest: ['RPSA', 'RPL23A', 'NUP93', 'AQR', 'RPA2', 'SUPT5H']\ntest: ['RPL6', 'RPS13', 'SF3B2', 'RPS27A', 'PRPF31', 'COPZ1']\ntest: ['RPS4X', 'PSMD1', 'RPS14', 'NUP98', 'USP39', 'CDC5L']\ntest: ['RPL5', 'PHB2', 'RPS15A', 'RPS3', 'ARCN1', 'COPS6']\ntest: ['RPS6', 'PRPF19', 'RPL34', 'Hsp10', 'POLR2I', 'EIF5B']\nauroc_mean: 0.8419599405820187\nauroc_sd: 0.01700381890902823\nauprc_mean: 0.5374780612257355\nauprc_sd: 0.0307901999759122\ntp_80_mean: 0.813415325970473\ntp_80_sd: 0.060331083529782635\ntp_90_mean: 0.8539377289377289\ntp_90_sd: 0.14257472730230836\n" ], [ "print(auroc_l)\nprint(auprc_l)\nprint(tp_80)\nprint(tp_90)", "[0.8584233873947144, 0.8178238022371683, 0.8167461147806936, 0.835408453266539, 0.8520322898053144, 0.8454747683961216, 0.8636994996978711, 0.8343741102400096, 0.8536570394197361]\n[0.566563028022413, 0.4848862901265541, 0.4920029175065841, 0.5422250027559219, 0.556203199805487, 0.5660378943500952, 0.5615582396576324, 0.5413416546584848, 0.5264843241484469]\n[0.8656716417910447, 0.676056338028169, 0.7966101694915254, 0.8414634146341463, 0.8536585365853658, 0.8666666666666667, 0.8387096774193549, 0.8041237113402062, 0.7777777777777778]\n[0.9285714285714286, 0.8461538461538461, 0.5833333333333334, 0.9285714285714286, 0.875, 1.0, 1.0, 0.8571428571428571, 0.6666666666666666]\n" ], [ "#GradientBoostingClassifier, hp2\nclf = ensemble.GradientBoostingClassifier(random_state=0,max_depth=4,\n max_features='sqrt', n_estimators=1800)\n\nauroc_l = []\nauprc_l = []\ntp_80 = []\ntp_90 = []\n\nfor s in range(9):\n tr, te = create_gene_splits_filter1_kfold_noval(dataframe['gene'].values, all_cols, 9, s)\n # training input and output\n d_input = tr[0]\n d_output = tr[1]\n clf.fit(d_input, d_output) #fit models\n #test set\n xt = te[0] \n pred = clf.predict_proba(xt)\n pred = pred[:,1]\n auroc,auprc,tp_thres = classification_analysis('GradientBoostingClassifier_hpnew', s,pred,te[1])\n auroc_l.append(auroc)\n auprc_l.append(auprc)\n if tp_thres[0]!= 'na':\n tp_80.append(tp_thres[0])\n if tp_thres[1]!= 'na':\n tp_90.append(tp_thres[1])\n \nauroc_mean = statistics.mean(auroc_l)\nauroc_sd = statistics.stdev(auroc_l)\nprint('auroc_mean: '+str(auroc_mean))\nprint('auroc_sd: '+str(auroc_sd))\nauprc_mean = statistics.mean(auprc_l)\nauprc_sd = statistics.stdev(auprc_l)\nprint('auprc_mean: '+str(auprc_mean))\nprint('auprc_sd: '+str(auprc_sd))\n \ntp_80_mean = statistics.mean(tp_80)\ntp_80_sd = statistics.stdev(tp_80)\nprint('tp_80_mean: '+str(tp_80_mean))\nprint('tp_80_sd: '+str(tp_80_sd))\ntp_90_mean = statistics.mean(tp_90)\ntp_90_sd = statistics.stdev(tp_90)\nprint('tp_90_mean: '+str(tp_90_mean))\nprint('tp_90_sd: '+str(tp_90_sd))", "test: ['RPL31', 'RPS3A', 'CSE1L', 'XAB2', 'PSMD7', 'SUPT6H']\ntest: ['EEF2', 'RPS11', 'SNRPD2', 'RPL37', 'SF3B3', 'DDX51']\ntest: ['RPL7', 'RPS9', 'KARS', 'SF3A1', 'RPL32', 'PSMB2']\ntest: ['RPS7', 'EIF4A3', 'U2AF1', 'PSMA1', 'PHB', 'POLR2D']\ntest: ['RPSA', 'RPL23A', 'NUP93', 'AQR', 'RPA2', 'SUPT5H']\ntest: ['RPL6', 'RPS13', 'SF3B2', 'RPS27A', 'PRPF31', 'COPZ1']\ntest: ['RPS4X', 'PSMD1', 'RPS14', 'NUP98', 'USP39', 'CDC5L']\ntest: ['RPL5', 'PHB2', 'RPS15A', 'RPS3', 'ARCN1', 'COPS6']\ntest: ['RPS6', 'PRPF19', 'RPL34', 'Hsp10', 'POLR2I', 'EIF5B']\nauroc_mean: 0.8402434698783054\nauroc_sd: 0.017096114410535924\nauprc_mean: 0.5326705713945947\nauprc_sd: 0.029089488378007556\ntp_80_mean: 0.8210448665312134\ntp_80_sd: 0.08843235627451937\ntp_90_mean: 0.8753086419753087\ntp_90_sd: 0.19982845866986979\n" ] ] ]

[ "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nfrom pathlib import Path ", "_____no_output_____" ], [ "from sklearn.utils import Bunch # Array, Target, Target names and so on", "_____no_output_____" ], [ "from skimage.io import imread # to read my images\nfrom skimage.transform import resize\nimport skimage as sk", "_____no_output_____" ], [ "def generate_data(Location , dim =(64,64)):\n \"\"\"We are Generating the data with different Dimensions of data set and we are bring them into 24*24\"\"\"\n \n img_dir = Path(Location)\n print(img_dir)\n folder_dir = [directory for directory in img_dir.iterdir()]\n print(folder_dir)\n Categories = [fo.name for fo in folder_dir]\n print(Categories)\n \n images = []\n image_data = []\n target = []\n \n desc = \"Image Generation\"\n \n for i, direc in enumerate(folder_dir):\n #print(i,direc)\n for files in direc.iterdir():\n #print(files)\n img = imread(files) \n img_resize = resize(img,dim)\n images.append(img_resize)\n image_data.append(img_resize.flatten())\n target.append(i)\n\n # Coverting list data into array Formate\n flat_data = np.array(image_data)\n target = np.array(target)\n images = np.array(images)\n return Bunch(data = flat_data,\n target= target,\n target_names = Categories,\n images =images,\n desc = desc)\n \n ", "_____no_output_____" ], [ "Dataset = generate_data(\"Datasets/Fruits\")", "Datasets\\Fruits\n[WindowsPath('Datasets/Fruits/Apples'), WindowsPath('Datasets/Fruits/Banana'), WindowsPath('Datasets/Fruits/Guava'), WindowsPath('Datasets/Fruits/Lemons')]\n['Apples', 'Banana', 'Guava', 'Lemons']\n" ], [ "Dataset.keys()", "_____no_output_____" ], [ "Dataset.data.shape", "_____no_output_____" ], [ "Dataset.target.shape", "_____no_output_____" ], [ "Dataset.target", "_____no_output_____" ], [ "sns.distplot(Dataset.data)", "_____no_output_____" ], [ "from sklearn.model_selection import train_test_split", "_____no_output_____" ], [ "x_train,x_test,y_train,y_test= train_test_split(Dataset.data,Dataset.target,test_size=0.20)", "_____no_output_____" ], [ "from sklearn.linear_model import LogisticRegression", "_____no_output_____" ], [ "Lr = LogisticRegression(max_iter=4000)", "_____no_output_____" ], [ "Lr.fit(x_train,y_train)", "_____no_output_____" ], [ "yhat = Lr.predict(x_test)\nyhat", "_____no_output_____" ], [ "pd.DataFrame({\"Actual\":y_test,\n \"Predications\":yhat,\n \"Same Data\":y_test==yhat})", "_____no_output_____" ], [ "from sklearn.metrics import confusion_matrix,classification_report,accuracy_score,plot_confusion_matrix", "_____no_output_____" ], [ "accuracy_score(y_test,yhat)", "_____no_output_____" ], [ "plot_confusion_matrix(Lr,x_test,y_test)", "_____no_output_____" ], [ "cm = confusion_matrix(y_test,yhat)\ncm", "_____no_output_____" ], [ "sns.heatmap(cm,annot=True)", "_____no_output_____" ], [ "print(classification_report(y_test,yhat))", " precision recall f1-score support\n\n 0 0.91 1.00 0.95 10\n 1 1.00 0.75 0.86 8\n 2 0.90 0.75 0.82 12\n 3 0.79 1.00 0.88 11\n\n accuracy 0.88 41\n macro avg 0.90 0.88 0.88 41\nweighted avg 0.89 0.88 0.88 41\n\n" ], [ "from sklearn.neighbors import KNeighborsClassifier", "_____no_output_____" ], [ "knn = KNeighborsClassifier(n_neighbors=2)", "_____no_output_____" ], [ "knn.fit(x_train,y_train)", "_____no_output_____" ], [ "knn_yhat = knn.predict(x_test)", "_____no_output_____" ], [ "D = pd.DataFrame({\"Logistics_New_predication\":yhat,\n \"Knn_New_predications\":knn_yhat,\n \"Actual\":y_test})", "_____no_output_____" ], [ "plot_confusion_matrix(knn,x_test,knn_yhat)", "_____no_output_____" ], [ "cm= confusion_matrix(y_test,knn_yhat)", "_____no_output_____" ], [ "sns.heatmap(cm,annot=True)", "_____no_output_____" ], [ "print(classification_report(y_test,knn_yhat))", " precision recall f1-score support\n\n 0 0.91 1.00 0.95 10\n 1 1.00 0.88 0.93 8\n 2 0.62 0.83 0.71 12\n 3 0.71 0.45 0.56 11\n\n accuracy 0.78 41\n macro avg 0.81 0.79 0.79 41\nweighted avg 0.79 0.78 0.77 41\n\n" ] ], [ [ "### Task:\n1. plot AUC or ROC ", "_____no_output_____" ] ] ]

[ "code", "markdown" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "empty" ] ] ]

[ "empty" ]

[ [ "empty" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Test functions", "_____no_output_____" ] ], [ [ "from utils.sparse import *", "_____no_output_____" ] ], [ [ "# Function list \n1. inplace_set_rows_zero_where_sum (X, op, cut) \n2. inplace_set_cols_zero_where_sum (X, op, cut)\n3. inplace_set_rows_zero (X, target_rows)\n4. inplace_set_cols_zero (X, target_cols)\n5. inplace_row_scale (X, scale)\n6. inplace_col_scale (X, scale) \n7. sum_cols (X)\n8. sum_rows (X)", "_____no_output_____" ] ], [ [ "m = sp.random(4,5,0.5).tocsr()\nm.data = np.ones(m.data.shape[0])\nprint(m.todense())\ninplace_row_scale(m,np.array([1,2,3,4]))\nprint (m.todense())", "[[0. 1. 1. 0. 0.]\n [1. 1. 0. 0. 0.]\n [0. 0. 1. 1. 1.]\n [1. 1. 1. 0. 0.]]\n[[0. 1. 1. 0. 0.]\n [2. 2. 0. 0. 0.]\n [0. 0. 3. 3. 3.]\n [4. 4. 4. 0. 0.]]\n" ], [ "m = sp.random(4,5,0.5).tocsc()\nm.data = np.ones(m.data.shape[0])\nprint(m.todense())\ninplace_col_scale(m,np.array([1,2,3,4,5]))\nprint (m.todense())", "[[1. 0. 0. 0. 0.]\n [0. 0. 1. 1. 1.]\n [0. 1. 0. 1. 1.]\n [1. 1. 0. 0. 1.]]\n[[1. 0. 0. 0. 0.]\n [0. 0. 3. 4. 5.]\n [0. 2. 0. 4. 5.]\n [1. 2. 0. 0. 5.]]\n" ], [ "m = sp.random(4,5,0.5).tocsr()\nm.data = np.ones(m.data.shape[0])\nprint(m.todense())\ninplace_set_rows_zero(m,np.array([1,3]))\nprint (m.todense())", "[[0. 0. 0. 0. 1.]\n [1. 0. 1. 1. 0.]\n [0. 0. 0. 1. 1.]\n [1. 1. 1. 0. 1.]]\n[[0. 0. 0. 0. 1.]\n [0. 0. 0. 0. 0.]\n [0. 0. 0. 1. 1.]\n [0. 0. 0. 0. 0.]]\n" ], [ "m = sp.random(4,5,0.5).tocsr()\nm.data = np.ones(m.data.shape[0])\nprint(m.todense())\ninplace_set_cols_zero(m,np.array([1,3]))\nprint (m.todense())", "[[0. 0. 1. 0. 1.]\n [1. 1. 0. 1. 0.]\n [0. 1. 0. 1. 0.]\n [0. 0. 1. 1. 1.]]\n[[0. 0. 1. 0. 1.]\n [1. 0. 0. 0. 0.]\n [0. 0. 0. 0. 0.]\n [0. 0. 1. 0. 1.]]\n" ], [ "m = sp.random(4,5,0.5).tocsr()\nprint (sum_rows(m))\ninplace_set_rows_zero_where_sum(m, '>', 1.5)\nprint (m.todense())", "[2.06284216 0.17633125 0.86179057 1.79197257]\n[[0. 0. 0. 0. 0. ]\n [0. 0. 0. 0. 0.17633125]\n [0.01446204 0. 0.39882688 0. 0.44850165]\n [0. 0. 0. 0. 0. ]]\n" ], [ "m = sp.random(4,5,0.5).tocsr()\nprint (sum_cols(m))\ninplace_set_cols_zero_where_sum(m, '>', 1.5)\nprint (m.todense())", "[1.96108189 1.12923879 0. 1.93997106 0.40970854]\n[[0. 0.69020914 0. 0. 0.40970854]\n [0. 0. 0. 0. 0. ]\n [0. 0. 0. 0. 0. ]\n [0. 0.43902965 0. 0. 0. ]]\n" ] ] ]

[ "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import matplotlib.pyplot as plt\nimport numpy as np\n\n#Código original do Rafael para carregar o dataset \n\ndef load(dataset):\n with open(dataset) as ds:\n d = []\n for l in ds.readlines():\n l = l.split()\n d.append([float(value) for value in l])\n return np.array(d)\n\ndef plot(x, y, b):\n # plota os pontos atuais \n # s = np.abs(y) esse valor pode ser útil para personalizar o Scatter Plot no futuro \n\n plt.scatter(x, y, c = x, cmap='plasma', marker = \"o\", s = 10)\n plt.plot(x, b, color='red')\n \n plt.axvline(0, c=(.5, .5, .5), ls = '--')\n plt.axhline(0, c=(.5, .5, .5), ls = '--')\n \n plt.colorbar()\n\n #legenda e título do gráfico\n plt.title('Modelos de regressão Linear')\n plt.xlabel('velocidade do vento – m/s')\n plt.ylabel('variável de saída: potência gerada – kWatts')\n plt.savefig('img/update.jpg', dpi=300)\n plt.show()\n \n \n\ndef main():\n base = load('aerogerador.dat')\n # separa as colunas do vetor\n x, y = base[:,0], base[:,1]\n n = len(y)\n\n # valor de beta_1 e beta_0\n beta_1 = (np.sum([x[i]*y[i] for i in range(0, len(x))]) - (1/n) * np.sum(y) * np.sum(x)) / (np.sum(x**2) - (1/n) * np.sum(x)**2)\n\n beta_0 = np.mean(y) - beta_1 * np.mean(x)\n\n # cria a equacao da reta\n y_chap = beta_0 + beta_1 * x\n\n\n print('Beta 1: {}\\nBeta 0: {}\\n'.format(beta_1, beta_0))\n print('\\n')\n\n plot(x,y,y_chap)\n\n\nif __name__ == \"__main__\":\n main() \n\n", "Beta 1: 56.443855448050236\nBeta 0: -217.69027909512027\n\n\n\n" ] ] ]

[ "code" ]

[ [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "## Pivot table\n- excel에서 보던 것\n- index축은 groupby와 동일\n- column에 추가로 labeling값을 추가하여,\n- Value에 numeric type 값을 aggregation하는 형태", "_____no_output_____" ] ], [ [ "import dateutil\n\ndf_phone = pd.read_csv(\"code/ch5/data/phone_data.csv\")\ndf_phone['date'] = df_phone['date'].apply(dateutil.parser.parse, dayfirst=True)\ndf_phone.tail()", "_____no_output_____" ], [ "df_phone.pivot_table(['duration'], index=['month','item'], columns=['network'], fill_value=0, aggfunc='sum')", "_____no_output_____" ] ], [ [ "## Crosstab\n- 두 컬럼의 교차 빈도, 비율, 덧셈 등을 구할 때 사용\n- Pivot table의 특수한 형태\n- User-Item Rating Matrix 등을 만들 때 사용가능", "_____no_output_____" ] ], [ [ "df_movie = pd.read_csv(\"code/ch5/data/movie_rating.csv\")\ndf_movie.tail()", "_____no_output_____" ], [ "# 평론가의 영화별 평점\npd.crosstab(values=df_movie.rating, index=df_movie.critic, columns=df_movie.title, aggfunc='first').fillna(0)", "_____no_output_____" ], [ "# 이걸 groupby로 만들어보자.1\ndf_movie.groupby(['critic','title'])['rating'].first().unstack().fillna(0)", "_____no_output_____" ], [ "# 이걸 groupby로 만들어보자.2\ndf_movie.groupby(['critic','title']).agg({'rating' : 'first'}).unstack().fillna(0)", "_____no_output_____" ], [ "# 이걸 pivot table로 만들어보자\ndf_movie.pivot_table(values='rating', index='critic', columns='title', aggfunc='first', fill_value=0)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ] ]

[ "CC-BY-4.0" ]

[ "CC-BY-4.0" ]

[ "CC-BY-4.0" ]

[ [ [ "from typing import Union, List\nimport networkx as nx\nimport pandas as pd\nimport math\nimport matplotlib.pyplot as plt\nfrom networkx.drawing.nx_agraph import graphviz_layout\nfrom abc import ABC, abstractmethod\nimport numpy as np", "_____no_output_____" ], [ "def down(n: int) -> int:\n if n % 24 == 5:\n return (n-1)/4\n elif n % 96 == 85:\n return (n-5)/16\n else:\n return -1\n\ndef right_prune1(n: int) -> Union[int, int]:\n d = down(n)\n mod_case = -1\n if (2*d)%18 == 4 and (d%18) in [11,17] :\n return 1, (8*d-1)/3\n elif (4*d)%18 == 4 and (d%18) in [11,13]:\n return 2, (16*d-1)/3\n elif (2*d)%18 == 16 and (d%18) in [11,17] or d%18 == 5:\n return 3, (32*d-1)/3\n elif (4*d)%18 == 16 and (d%18) in [1,13] or d%18 == 7:\n return 4, (64*d-1)/3\n else:\n return 0, -1\n\ndef right_prune2(n: int) -> Union[int, int]:\n dd = down(down(n))\n mod_case = -1\n if (2*d)%18 == 4 and (d%18) in [11,17] :\n return 1, (32*dd-1)/3\n elif (4*d)%18 == 4 and (d%18) in [11,13]:\n return 2, (64*dd-1)/3\n elif (2*d)%18 == 16 and (d%18) in [11,17] or d%18 == 5:\n return 3, (128*dd-1)/3\n elif (4*d)%18 == 16 and (d%18) in [1,13] or d%18 == 7:\n return 4, (256*dd-1)/3\n elif (4*d)%18 == 16 and (d%18) in [1,13] or d%18 == 7:\n return 4, (512*dd-1)/3\n elif (4*d)%18 == 16 and (d%18) in [1,13] or d%18 == 7:\n return 4, (1024*dd-1)/3\n else:\n return 0, -1", "_____no_output_____" ], [ "#print(down(53))#13\n#print(down(1109))#277\n#print(down(181))#11\n\n#case 2 never happens\nprint(right_prune1(85))#53\nprint(right_prune1(341))#1813\nprint(right_prune1(181))#29\nprint(right_prune1(1813))#1205\nprint(right_prune1(53))#1205\n", "(3, 53.0)\n(4, 1813.0)\n(1, 29.0)\n(3, 1205.0)\n(4, 277.0)\n" ] ] ]

[ "code" ]

[ [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "## MNIST Simple DEMO", "_____no_output_____" ] ], [ [ "import argparse\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nimport torch.optim as optim\nfrom torchvision import datasets, transforms", "_____no_output_____" ], [ "class Arguments:\n batch = 64\n test_batch = 512\n epochs = 10\n lr = .01\n momentum = .5\n seed = 42\n log_interval = 100", "_____no_output_____" ], [ "args = Arguments()", "_____no_output_____" ], [ "class Network(nn.Module):\n def __init__(self):\n super(Network, self).__init__()\n self.conv1 = nn.Conv2d(1, 20, 5, 1)\n self.conv2 = nn.Conv2d(20, 50, 5, 1)\n self.fc1 = nn.Linear(4*4*50, 500)\n self.fc2 = nn.Linear(500, 10)\n\n def forward(self, x):\n x = F.relu(self.conv1(x))\n x = F.max_pool2d(x, 2, 2)\n x = F.relu(self.conv2(x))\n x = F.max_pool2d(x, 2, 2)\n x = x.view(-1, 4*4*50)\n x = F.relu(self.fc1(x))\n x = self.fc2(x)\n return F.log_softmax(x, dim=1)", "_____no_output_____" ], [ "def train(args, model, device, train_loader, optimizer, epoch):\n model.train()\n\n for batch_idx, (data, target) in enumerate(train_loader):\n data, target = data.to(device), target.to(device)\n\n optimizer.zero_grad()\n output = model(data)\n loss = F.nll_loss(output, target)\n loss.backward()\n optimizer.step()\n\n if batch_idx % args.log_interval == 0:\n print('Train Epoch: {} [{}/{} ({:.0f}%)]\\tLoss: {:.6f}'.format(\n epoch, batch_idx * len(data), len(train_loader.dataset),\n 100. * batch_idx / len(train_loader), loss.item()))\n\n\ndef test(args, model, device, test_loader):\n model.eval()\n test_loss, correct = 0, 0\n\n with torch.no_grad():\n for data, target in test_loader:\n data, target = data.to(device), target.to(device)\n\n output = model(data)\n test_loss += F.nll_loss(output, target, reduction='sum').item()\n pred = output.argmax(dim=1, keepdim=True)\n correct += pred.eq(target.view_as(pred)).sum().item()\n\n test_loss /= len(test_loader.dataset)\n\n print('\\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\\n'.format(\n test_loss, correct, len(test_loader.dataset),\n 100. * correct / len(test_loader.dataset)))", "_____no_output_____" ], [ "torch.manual_seed(args.seed)\ndevice = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\nkwargs = {\n 'num_workers': 1,\n 'pin_memory': True\n} if device.type == 'cuda' else {}\n\ntrain_loader = torch.utils.data.DataLoader(\n datasets.MNIST('../data', train=True, download=True,\n transform=transforms.Compose([\n transforms.ToTensor(),\n transforms.Normalize((0.1307,), (0.3081,))\n ])),\n batch_size=args.batch, shuffle=True, **kwargs)\ntest_loader = torch.utils.data.DataLoader(\n datasets.MNIST('../data', train=False, transform=transforms.Compose([\n transforms.ToTensor(),\n transforms.Normalize((0.1307,), (0.3081,))\n ])),\n batch_size=args.test_batch, shuffle=True, **kwargs)\n\nmodel = Network().to(device)\noptimizer = optim.SGD(model.parameters(), lr=args.lr, momentum=args.momentum)\n\nfor epoch in range(1, args.epochs + 1):\n train(args, model, device, train_loader, optimizer, epoch)\n test(args, model, device, test_loader)\n\ntorch.save(model.state_dict(), \"mnist_cnn.pt\")", "Train Epoch: 1 [0/60000 (0%)]\tLoss: 2.309220\nTrain Epoch: 1 [6400/60000 (11%)]\tLoss: 0.545335\nTrain Epoch: 1 [12800/60000 (21%)]\tLoss: 0.417650\nTrain Epoch: 1 [19200/60000 (32%)]\tLoss: 0.353491\nTrain Epoch: 1 [25600/60000 (43%)]\tLoss: 0.306972\nTrain Epoch: 1 [32000/60000 (53%)]\tLoss: 0.133229\nTrain Epoch: 1 [38400/60000 (64%)]\tLoss: 0.188936\nTrain Epoch: 1 [44800/60000 (75%)]\tLoss: 0.070623\nTrain Epoch: 1 [51200/60000 (85%)]\tLoss: 0.258176\nTrain Epoch: 1 [57600/60000 (96%)]\tLoss: 0.040762\n\nTest set: Average loss: 0.1040, Accuracy: 9675/10000 (97%)\n\nTrain Epoch: 2 [0/60000 (0%)]\tLoss: 0.235796\nTrain Epoch: 2 [6400/60000 (11%)]\tLoss: 0.049525\nTrain Epoch: 2 [12800/60000 (21%)]\tLoss: 0.077299\nTrain Epoch: 2 [19200/60000 (32%)]\tLoss: 0.058649\nTrain Epoch: 2 [25600/60000 (43%)]\tLoss: 0.162579\nTrain Epoch: 2 [32000/60000 (53%)]\tLoss: 0.043902\nTrain Epoch: 2 [38400/60000 (64%)]\tLoss: 0.037764\nTrain Epoch: 2 [44800/60000 (75%)]\tLoss: 0.007759\nTrain Epoch: 2 [51200/60000 (85%)]\tLoss: 0.125971\nTrain Epoch: 2 [57600/60000 (96%)]\tLoss: 0.033037\n\nTest set: Average loss: 0.0616, Accuracy: 9805/10000 (98%)\n\nTrain Epoch: 3 [0/60000 (0%)]\tLoss: 0.081351\nTrain Epoch: 3 [6400/60000 (11%)]\tLoss: 0.088761\nTrain Epoch: 3 [12800/60000 (21%)]\tLoss: 0.095073\nTrain Epoch: 3 [19200/60000 (32%)]\tLoss: 0.091261\nTrain Epoch: 3 [25600/60000 (43%)]\tLoss: 0.160844\nTrain Epoch: 3 [32000/60000 (53%)]\tLoss: 0.034395\nTrain Epoch: 3 [38400/60000 (64%)]\tLoss: 0.010957\nTrain Epoch: 3 [44800/60000 (75%)]\tLoss: 0.033368\nTrain Epoch: 3 [51200/60000 (85%)]\tLoss: 0.013109\nTrain Epoch: 3 [57600/60000 (96%)]\tLoss: 0.070705\n\nTest set: Average loss: 0.0484, Accuracy: 9847/10000 (98%)\n\nTrain Epoch: 4 [0/60000 (0%)]\tLoss: 0.019743\nTrain Epoch: 4 [6400/60000 (11%)]\tLoss: 0.040987\nTrain Epoch: 4 [12800/60000 (21%)]\tLoss: 0.061202\nTrain Epoch: 4 [19200/60000 (32%)]\tLoss: 0.007646\nTrain Epoch: 4 [25600/60000 (43%)]\tLoss: 0.011820\nTrain Epoch: 4 [32000/60000 (53%)]\tLoss: 0.022924\nTrain Epoch: 4 [38400/60000 (64%)]\tLoss: 0.044619\nTrain Epoch: 4 [44800/60000 (75%)]\tLoss: 0.015211\nTrain Epoch: 4 [51200/60000 (85%)]\tLoss: 0.016549\nTrain Epoch: 4 [57600/60000 (96%)]\tLoss: 0.069062\n\nTest set: Average loss: 0.0358, Accuracy: 9887/10000 (99%)\n\nTrain Epoch: 5 [0/60000 (0%)]\tLoss: 0.036325\nTrain Epoch: 5 [6400/60000 (11%)]\tLoss: 0.068640\nTrain Epoch: 5 [12800/60000 (21%)]\tLoss: 0.010548\nTrain Epoch: 5 [19200/60000 (32%)]\tLoss: 0.029485\nTrain Epoch: 5 [25600/60000 (43%)]\tLoss: 0.025582\nTrain Epoch: 5 [32000/60000 (53%)]\tLoss: 0.060043\nTrain Epoch: 5 [38400/60000 (64%)]\tLoss: 0.013400\nTrain Epoch: 5 [44800/60000 (75%)]\tLoss: 0.011863\nTrain Epoch: 5 [51200/60000 (85%)]\tLoss: 0.067035\nTrain Epoch: 5 [57600/60000 (96%)]\tLoss: 0.056927\n\nTest set: Average loss: 0.0344, Accuracy: 9884/10000 (99%)\n\nTrain Epoch: 6 [0/60000 (0%)]\tLoss: 0.014376\nTrain Epoch: 6 [6400/60000 (11%)]\tLoss: 0.006622\nTrain Epoch: 6 [12800/60000 (21%)]\tLoss: 0.020543\nTrain Epoch: 6 [19200/60000 (32%)]\tLoss: 0.035187\nTrain Epoch: 6 [25600/60000 (43%)]\tLoss: 0.038597\nTrain Epoch: 6 [32000/60000 (53%)]\tLoss: 0.016477\nTrain Epoch: 6 [38400/60000 (64%)]\tLoss: 0.021265\nTrain Epoch: 6 [44800/60000 (75%)]\tLoss: 0.034409\nTrain Epoch: 6 [51200/60000 (85%)]\tLoss: 0.012662\nTrain Epoch: 6 [57600/60000 (96%)]\tLoss: 0.044574\n\nTest set: Average loss: 0.0375, Accuracy: 9879/10000 (99%)\n\nTrain Epoch: 7 [0/60000 (0%)]\tLoss: 0.011418\nTrain Epoch: 7 [6400/60000 (11%)]\tLoss: 0.008460\nTrain Epoch: 7 [12800/60000 (21%)]\tLoss: 0.024678\nTrain Epoch: 7 [19200/60000 (32%)]\tLoss: 0.021109\nTrain Epoch: 7 [25600/60000 (43%)]\tLoss: 0.044059\nTrain Epoch: 7 [32000/60000 (53%)]\tLoss: 0.012801\nTrain Epoch: 7 [38400/60000 (64%)]\tLoss: 0.002572\nTrain Epoch: 7 [44800/60000 (75%)]\tLoss: 0.008726\nTrain Epoch: 7 [51200/60000 (85%)]\tLoss: 0.032433\nTrain Epoch: 7 [57600/60000 (96%)]\tLoss: 0.086093\n\nTest set: Average loss: 0.0300, Accuracy: 9900/10000 (99%)\n\nTrain Epoch: 8 [0/60000 (0%)]\tLoss: 0.005734\nTrain Epoch: 8 [6400/60000 (11%)]\tLoss: 0.011664\nTrain Epoch: 8 [12800/60000 (21%)]\tLoss: 0.083290\nTrain Epoch: 8 [19200/60000 (32%)]\tLoss: 0.014290\nTrain Epoch: 8 [25600/60000 (43%)]\tLoss: 0.018174\nTrain Epoch: 8 [32000/60000 (53%)]\tLoss: 0.013148\nTrain Epoch: 8 [38400/60000 (64%)]\tLoss: 0.010231\nTrain Epoch: 8 [44800/60000 (75%)]\tLoss: 0.054055\nTrain Epoch: 8 [51200/60000 (85%)]\tLoss: 0.003165\nTrain Epoch: 8 [57600/60000 (96%)]\tLoss: 0.023597\n\nTest set: Average loss: 0.0319, Accuracy: 9884/10000 (99%)\n\nTrain Epoch: 9 [0/60000 (0%)]\tLoss: 0.056386\nTrain Epoch: 9 [6400/60000 (11%)]\tLoss: 0.022121\nTrain Epoch: 9 [12800/60000 (21%)]\tLoss: 0.024276\nTrain Epoch: 9 [19200/60000 (32%)]\tLoss: 0.014277\nTrain Epoch: 9 [25600/60000 (43%)]\tLoss: 0.027978\nTrain Epoch: 9 [32000/60000 (53%)]\tLoss: 0.007992\nTrain Epoch: 9 [38400/60000 (64%)]\tLoss: 0.018210\nTrain Epoch: 9 [44800/60000 (75%)]\tLoss: 0.023663\nTrain Epoch: 9 [51200/60000 (85%)]\tLoss: 0.005544\nTrain Epoch: 9 [57600/60000 (96%)]\tLoss: 0.005737\n\nTest set: Average loss: 0.0281, Accuracy: 9906/10000 (99%)\n\nTrain Epoch: 10 [0/60000 (0%)]\tLoss: 0.011280\nTrain Epoch: 10 [6400/60000 (11%)]\tLoss: 0.029055\nTrain Epoch: 10 [12800/60000 (21%)]\tLoss: 0.007866\nTrain Epoch: 10 [19200/60000 (32%)]\tLoss: 0.053182\nTrain Epoch: 10 [25600/60000 (43%)]\tLoss: 0.002478\nTrain Epoch: 10 [32000/60000 (53%)]\tLoss: 0.001874\nTrain Epoch: 10 [38400/60000 (64%)]\tLoss: 0.041121\nTrain Epoch: 10 [44800/60000 (75%)]\tLoss: 0.004530\nTrain Epoch: 10 [51200/60000 (85%)]\tLoss: 0.038643\nTrain Epoch: 10 [57600/60000 (96%)]\tLoss: 0.008336\n\nTest set: Average loss: 0.0264, Accuracy: 9910/10000 (99%)\n\n" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "import simplejson as json\n\nwith open('0001_bytearray_master.json') as f:\n byte_array_data = json.load(f)\n\nwith open('0002_binary_master.json') as f:\n bin_obj_data = json.load(f)", "_____no_output_____" ], [ "from collections import defaultdict\n\n\ndef prepare_data(data, name_pattern='', max_threshold=102400, min_threshold=0):\n ret = {'simple': defaultdict(dict), 'partition_aware': defaultdict(dict)}\n\n for bench in data['benchmarks']:\n params = bench['param'].split('-')\n name = bench['name']\n if name_pattern not in name:\n continue\n \n batch_sz = int(params[2]) if len(params) == 3 else 0\n value_sz = int(params[1])\n\n if not (min_threshold < value_sz < max_threshold):\n continue\n \n if 'async' in name:\n continue\n \n median = bench['stats']['median'] * 1000000\n ret[params[0]][value_sz] = median\n return ret", "_____no_output_____" ], [ "import matplotlib\nimport matplotlib.pyplot as plt\nimport numpy as np\n\n\ndef plot(pa_data, simple_data, title='',yscale='linear', time_scale=1, ylabel='us'):\n labels = []\n for k in pa_data.keys():\n if k // (1024 * 1024) > 0 or k == 1024 * 1024:\n labels.append(f'{k / (1024 * 1024)}Mb')\n elif k // 1024 > 0:\n labels.append(f'{k / 1024}Kb')\n else:\n labels.append(k)\n\n \n pa_means = [v / time_scale for v in pa_data.values()]\n simple_means = [v / time_scale for v in simple_data.values()]\n \n\n x = np.arange(len(labels))\n width = 0.3\n\n fig, ax = plt.subplots()\n plt.rcParams[\"figure.figsize\"] = (15, 12)\n plt.yscale(yscale)\n \n rects = []\n \n rects.append(ax.bar(x - width * 0.5, simple_means, width, label='simple'))\n rects.append(ax.bar(x + width * 0.5, pa_means, width, label='partition_aware'))\n\n ax.set_title(title)\n ax.set_ylabel(ylabel)\n ax.set_xticks(x)\n ax.set_xticklabels(labels)\n ax.legend()\n\n for rect in rects:\n ax.bar_label(rect, padding=3)\n\n\n fig.tight_layout()\n\n plt.show()", "_____no_output_____" ], [ "get_data = prepare_data(byte_array_data, 'get', max_threshold=20 * 1024 * 1024)\n\nplot(get_data['partition_aware'], get_data['simple'], 'Get bytearrays', yscale='log', ylabel='ms', time_scale=1000)", "_____no_output_____" ], [ "put_data = prepare_data(byte_array_data, 'put', max_threshold=20 * 1024 * 1024)\n\nplot(put_data['partition_aware'], put_data['simple'], 'Put bytearrays', yscale='log', ylabel='ms', time_scale=1000)", "_____no_output_____" ], [ "get_data = prepare_data(bin_obj_data, 'get', max_threshold=20 * 1024 * 1024)\n\nplot(get_data['partition_aware'], get_data['simple'], 'Get BinaryObject', yscale='log', ylabel='ms', time_scale=1000)", "_____no_output_____" ], [ "put_data = prepare_data(bin_obj_data, 'put', max_threshold=20 * 1024 * 1024)\n\nplot(put_data['partition_aware'], put_data['simple'], 'Put BinaryObject', yscale='log', ylabel='ms', time_scale=1000)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<h1>Project 0: Inaugural project </h1>", "_____no_output_____" ], [ "\n<b>Labor Supply Problem</b>", "_____no_output_____" ], [ "Following labor supply problem is given: \n\n$$\nc^*,l^* = log(c) - v \\frac{l^{1+\\frac{1}{\\epsilon}}}{1+\\frac{1}{\\epsilon}}\n\\\\\nx = m + wl - [\\tau_0wl+\\tau_1 \\max(wl-\\kappa,0)]\\\\\n\nc \\in [0,x]\\\\\nl \\in [0,1]\\\\\n\n$$\n\nWhere: \nc is consumption,\nl is labor supply,\nm is cash-on-hand, \nw is the wage rate, \n$$t_0$$ is the standard labor income tax\n$$t_1$$ is the top bracket labor income tax,\nk is the cut-off of top labor income bracket\nx is total resources\nv scales disutility of labor \nE is the Frisch elasticity of labor supply\nutility is monotonically increasing in consumption, which implies $$c^* = x$$", "_____no_output_____" ], [ "<h2>Question 1</h2>", "_____no_output_____" ] ], [ [ "# All used packages are imported\n\nimport numpy as np\nimport sympy as sm \nfrom scipy import optimize\n\n", "_____no_output_____" ], [ "t0 = sm.symbols('t_0')\nt1 = sm.symbols('t_1')", "_____no_output_____" ], [ "\n\n", "_____no_output_____" ], [ "m = 1 #cash-on-hand\nv = 10 #disutility of labor\ne = 0.3 #elasticity of labor supply\nt0 = 0.4 #standard labor income tax\nt1 = 0.1 #top bracket labor income tax\nk = 0.4 #cut-off for top labor income tax\n\n\n# Defining utility\n\ndef utility(c,v,l,e):\n u = np.log(c) - v*(l**(1+1/e)/(1+1/e))\n return u\n\n# Defining constraint\n\ndef constraint(m,w,l,t0,t1,k):\n x = m + w*l - (t0*w*l + t1*np.max(w*l-k,0))\n return x\n\n\ndef consumption(l,w,e,v,t0,t1,k):\n c = constraint(m,w,l,t0,t1,k)\n return -utility(c,v,l,e)\n\n\ndef optimizer(w,e,v,t0,t1,k,m):\n res = optimize.minimize_scalar(\n consumption, method='bounded',\n bounds=(0,1), args=(w,e,v,t0,t1,k))\n \n labor_star = res.x\n cons_star = constraint(m,w,labor_star,t0,t1,k)\n utility_star = utility(cons_star,v,labor_star,e)\n \n return labor_star,cons_star,utility_star\n\nlabor_star = optimizer(0.5,e,v,t0,t1,k,m)[0]\ncons_star = optimizer(0.5,e,v,t0,t1,k,m)[1]\nu_star = optimizer(0.5,e,v,t0,t1,k,m)[2]\n\nprint('labour supply is:' + str(labor_star))\nprint('consumption is:' + str(cons_star))\nprint('utility:' + str(u_star))", "labour supply is:0.31961536193545265\nconsumption is:1.119903840483863\nutility:0.09677772523865749\n" ] ], [ [ "<h2>Question 2</h2>", "_____no_output_____" ] ], [ [ "import matplotlib.pyplot as plt\n\n\nplt.style.use('grayscale')\n\n# Plot l_star and c_star with w going from 0.5 to 1.5\n# The definitions are defined - the used packages is defined above\nN = 10000\nw_vector = np.linspace(0.5,1.5,num=N)\nc_optimal = np.empty(N)\nl_optimal = np.empty(N)\n\n# a loop is generated to test the range of W \n\nfor i, w in enumerate(w_vector):\n optimization = optimizer(w,e,v,t0,t1,k,m)\n l_optimal[i]=optimization[0]\n c_optimal[i]=optimization[1]\n\nfig = plt.figure(figsize=(10,4))\n\n# Left plot\naxis_left = fig.add_subplot(1,2,1)\naxis_left.plot(w_vector,l_optimal)\naxis_left.set_title('Optimal labor supply given w')\naxis_left.set_xlabel('$w$')\naxis_left.set_ylabel('$l$')\naxis_left.grid(True)\n\n# Right plot \naxis_right = fig.add_subplot(1,2,2)\naxis_right.plot(w_vector,c_optimal)\naxis_right.set_title('Optimal consumption given w')\naxis_right.set_xlabel('$w1$')\naxis_right.set_ylabel('$c$')\naxis_right.grid(True)\n\nplt.show\n", "_____no_output_____" ] ], [ [ "\n<h2>Question 3</h2>", "_____no_output_____" ] ], [ [ "# Calculate the tax revenue\n\ntax_revenue = np.sum( t0 * w_vector * l_optimal + t1 * np.max( w_vector * l_optimal - k ,0 ))\nprint('Total tax revenue is: ' + str(tax_revenue))\n", "Total tax revenue is: 1775.3896759006836\n" ] ], [ [ "\n<h2>Question 4</h2>", "_____no_output_____" ] ], [ [ "# How does the tax revenue change when e = 0.1? \n# New epsilon is defined\ne_new = 0.1\nl_optimal_e_new = np.empty(N)\n\n# Same loop is used as above but only a new labor\n# supply is calculated as consumption isn't included\n# in the tax revenue formula\nfor i, w in enumerate(w_vector):\n optimization = optimizer(w,e_new,v,t0,t1,k,m)\n l_optimal_e_new[i]=optimization[0]\n\n# then the new tax revenue can be calculated\ntax_revenue_e_new = np.sum( t0 * w_vector * l_optimal_e_new + t1 * np.max( w_vector * l_optimal_e_new - k ,0))\nprint('New total tax revenue: '+str(tax_revenue_e_new))\n\n# Thus the difference in tax revenue can be calucalted as\nprint('The difference in tax revenue is: '+ str(tax_revenue_e_new-tax_revenue))", "New total tax revenue: 3578.900497991557\nThe difference in tax revenue is: 1803.5108220908735\n" ] ], [ [ "\n<h2>Question 5</h2>", "_____no_output_____" ] ], [ [ "# Optimize the tax \n\n# Same optimization formula as above\ndef tax_optimize(t0,t1,k):\n tax_optimal = optimize.minimize_scalar(tax_revenue , method='bounded' , x=[0.1,0.1,0.1])\n t0_optimal = tax_optimal.x\n t1_optimal = tax_optimal.x\n k_optimal = tax_optimal.x\n return t0_optimal, t1_optimal, k_optimal\n \nt0_optimal = tax_optimize(t0,t1,k)[0]\nt1_optimal = tax_optimize(t0,t1,k)[1]\nk_optimal = tax_optimize(t0,t1,k)[2]\n\nprint('Optimal t0 is: ' + str(t0_optimal))", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown", "markdown", "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "## Tic-Tac-Toe Agent\n​\nIn this notebook, you will learn to build an RL agent (using Q-learning) that learns to play Numerical Tic-Tac-Toe with odd numbers. The environment is playing randomly with the agent, i.e. its strategy is to put an even number randomly in an empty cell. The following is the layout of the notebook:\n - Defining epsilon-greedy strategy\n - Tracking state-action pairs for convergence\n - Define hyperparameters for the Q-learning algorithm\n - Generating episode and applying Q-update equation\n - Checking convergence in Q-values", "_____no_output_____" ], [ "#### Importing libraries\nWrite the code to import Tic-Tac-Toe class from the environment file", "_____no_output_____" ] ], [ [ "# from <TC_Env> import <TicTacToe> - import your class from environment file\nfrom TCGame_Env import TicTacToe\nimport collections\nimport numpy as np\nimport random\nimport pickle\nimport time\nfrom matplotlib import pyplot as plt\nfrom tqdm import tqdm", "_____no_output_____" ], [ "# Function to convert state array into a string to store it as keys in the dictionary\n# states in Q-dictionary will be of form: x-4-5-3-8-x-x-x-x\n# x | 4 | 5\n# ----------\n# 3 | 8 | x\n# ----------\n# x | x | x\n\ndef Q_state(state):\n\n return ('-'.join(str(e) for e in state)).replace('nan','x')", "_____no_output_____" ], [ "# Defining a function which will return valid (all possible actions) actions corresponding to a state\n# Important to avoid errors during deployment.\n\ndef valid_actions(state):\n\n valid_Actions = []\n \n valid_Actions = [i for i in env.action_space(state)[0]] ###### -------please call your environment as env\n return valid_Actions", "_____no_output_____" ], [ "# Defining a function which will add new Q-values to the Q-dictionary. \ndef add_to_dict(state):\n state1 = Q_state(state)\n \n valid_act = valid_actions(state)\n if state1 not in Q_dict.keys():\n for action in valid_act:\n Q_dict[state1][action]=0", "_____no_output_____" ] ], [ [ "#### Epsilon-greedy strategy - Write your code here\n\n(you can build your epsilon-decay function similar to the one given at the end of the notebook)", "_____no_output_____" ] ], [ [ "# Defining epsilon-greedy policy. You can choose any function epsilon-decay strategy\ndef epsilon_greedy(state, time):\n max_epsilon = 1.0\n min_epsilon = 0.001\n\n epsilon = min_epsilon + (max_epsilon - min_epsilon) * np.exp(-0.000001*time)\n z = np.random.random() \n if z > epsilon:\n action = max(Q_dict[Q_state(state)],key=Q_dict[Q_state(state)].get)\n else:\n action = random.sample(valid_actions(state),1)[0] \n \n return action", "_____no_output_____" ] ], [ [ "#### Tracking the state-action pairs for checking convergence - write your code here", "_____no_output_____" ] ], [ [ "# Initialise Q_dictionary as 'Q_dict' and States_tracked as 'States_track' (for convergence)\nQ_dict = collections.defaultdict(dict)\n\nStates_track = collections.defaultdict(dict) \n \nprint(len(Q_dict))\nprint(len(States_track))", "_____no_output_____" ], [ "# Initialise states to be tracked\ndef initialise_tracking_states():\n sample_q_values = [('x-3-x-x-x-6-x-x-x',(0,1)),\n ('x-1-x-x-x-x-8-x-x',(2,9)),\n ('x-x-x-x-6-x-x-x-5',(2,7)),\n ('x-x-x-x-9-x-6-x-x',(1,7)),\n ('x-5-x-2-x-x-4-7-x',(0,9)),\n ('9-x-5-x-x-x-8-x-4',(1,3)),\n ('2-7-x-x-6-x-x-3-x',(8,5)),\n ('9-x-x-x-x-2-x-x-x',(2,5)),\n ('x-x-7-x-x-x-x-x-2',(1,5)),\n ('5-x-x-x-x-6-x-x-x',(4,9)),\n ('4-x-x-6-x-x-3-1-x',(8,5)),\n ('5-x-8-x-x-6-3-x-x',(3,1)),\n ('x-6-5-x-2-x-x-3-x',(0,7)),\n ('7-x-5-x-2-x-x-x-6',(1,3))]\n \n for q_values in sample_q_values:\n state = q_values[0]\n action = q_values[1]\n States_track[state][action] = []", "_____no_output_____" ], [ "#Defining a function to save the Q-dictionary as a pickle file\n\ndef save_obj(obj, name ):\n with open(name + '.pkl', 'wb') as f:\n pickle.dump(obj, f, pickle.HIGHEST_PROTOCOL)", "_____no_output_____" ], [ "def save_tracking_states():\n for state in States_track.keys():\n for action in States_track[state].keys():\n if state in Q_dict and action in Q_dict[state]:\n States_track[state][action].append(Q_dict[state][action])", "_____no_output_____" ], [ "initialise_tracking_states()", "_____no_output_____" ] ], [ [ "#### Define hyperparameters ---write your code here", "_____no_output_____" ] ], [ [ "EPISODES = 6000000\nLR = 0.20\nGAMMA = 0.8\nthreshold = 2540\ncheckpoint_print_episodes = 600000", "_____no_output_____" ] ], [ [ "### Q-update loop ---write your code here", "_____no_output_____" ] ], [ [ "start_time = time.time()\n\nq_track={}\nq_track['x-3-x-x-x-6-x-x-x']=[]\nq_track['x-1-x-x-x-x-8-x-x']=[]\nq_track['x-x-x-x-6-x-x-x-5']=[]\nq_track['x-x-x-x-9-x-6-x-x']=[]\nq_track['x-5-x-2-x-x-4-7-x']=[]\nq_track['9-x-5-x-x-x-8-x-4']=[]\nq_track['2-7-x-x-6-x-x-3-x']=[]\nq_track['9-x-x-x-x-2-x-x-x']=[]\nq_track['x-x-7-x-x-x-x-x-2']=[]\nq_track['5-x-x-x-x-6-x-x-x']=[]\nq_track['4-x-x-6-x-x-3-1-x']=[]\nq_track['5-x-8-x-x-6-3-x-x']=[]\nq_track['x-6-5-x-2-x-x-3-x']=[]\nq_track['7-x-5-x-2-x-x-x-6']=[] \n \nagent_won_count = 0\nenv_won_count = 0\ntie_count = 0\n\nfor episode in range(EPISODES):\n ##### Start writing your code from the next line\n env = TicTacToe()\n \n ## Initalizing parameter for the episodes\n reward=0\n curr_state = env.state\n add_to_dict(curr_state)\n is_terminal = False\n total_reward = 0\n \n while not(is_terminal):\n curr_action = epsilon_greedy(curr_state, episode)\n \n if Q_state(curr_state) in q_track.keys():\n q_track[Q_state(curr_state)].append(curr_action)\n\n next_state,reward,is_terminal, msg = env.step(curr_state,curr_action) \n\n curr_lookup = Q_state(curr_state)\n next_lookup = Q_state(next_state)\n\n if is_terminal:\n q_value_max = 0\n \n # Tracking the count of games won by agent and environment\n if msg == \"Agent Won!\":\n agent_won_count += 1\n elif msg == \"Environment Won!\":\n env_won_count += 1\n else:\n tie_count += 1\n else:\n add_to_dict(next_state)\n max_next = max(Q_dict[next_lookup],key=Q_dict[next_lookup].get)\n q_value_max = Q_dict[next_lookup][max_next]\n\n Q_dict[curr_lookup][curr_action] += LR * ((reward + (GAMMA * (q_value_max))) - Q_dict[curr_lookup][curr_action]) \n curr_state = next_state\n\n total_reward += reward\n\n if (episode + 1) % checkpoint_print_episodes == 0:\n print(\"After playing %d games, Agent Won : %.4f, Environment Won : %.4f, Tie : %.4f\"% (episode + 1, \n agent_won_count / (episode + 1), env_won_count /(episode + 1), tie_count / (episode + 1)))\n\n if ((episode + 1) % threshold) == 0: \n save_tracking_states()\n\n if ((episode + 1) % 1000000) == 0:\n print('Processed %dM episodes'%((episode+1)/1000000))\n \nelapsed_time = time.time() - start_time\nsave_obj(States_track,'States_tracked') \nsave_obj(Q_dict,'Policy')", "_____no_output_____" ] ], [ [ "#### Check the Q-dictionary", "_____no_output_____" ] ], [ [ "Q_dict", "_____no_output_____" ], [ "len(Q_dict)", "_____no_output_____" ], [ "# try checking for one of the states - that which action your agent thinks is the best -----This will not be evaluated\nQ_dict['x-x-5-x-x-x-x-x-4']", "_____no_output_____" ] ], [ [ "#### Check the states tracked for Q-values convergence\n(non-evaluative)", "_____no_output_____" ] ], [ [ "# Write the code for plotting the graphs for state-action pairs tracked", "_____no_output_____" ], [ "plt.figure(0, figsize=(16,7))\nplt.subplot(241)\nt1=States_track['x-3-x-x-x-6-x-x-x'][(0,1)]\nplt.title(\"(s,a)=('x-3-x-x-x-6-x-x-x',(0,1))\")\nplt.plot(np.asarray(range(0, len(t1))),np.asarray(t1))\n\nplt.subplot(242)\nt2=States_track['x-x-x-x-6-x-x-x-5'][(2,7)]\nplt.title(\"(s,a)=('x-x-x-x-6-x-x-x-5',(2,7))\")\nplt.plot(np.asarray(range(0, len(t2))),np.asarray(t2))\n\nplt.subplot(243)\nt3=States_track['5-x-x-x-x-6-x-x-x'][(4,9)]\nplt.title(\"(s,a)=('5-x-x-x-x-6-x-x-x',(4,9))\")\nplt.plot(np.asarray(range(0, len(t3))),np.asarray(t3))\n\nplt.subplot(244)\nt4=States_track['x-5-x-2-x-x-4-7-x'][(0,9)]\nplt.title(\"(s,a)=('x-5-x-2-x-x-4-7-x',(0,9))\")\nplt.plot(np.asarray(range(0, len(t4))),np.asarray(t4))\n\nplt.show()", "_____no_output_____" ] ], [ [ "### Epsilon - decay check", "_____no_output_____" ] ], [ [ "max_epsilon = 1.0\nmin_epsilon = 0.001\ntime = np.arange(0,5000000)\nepsilon = []\nfor i in range(0,5000000):\n epsilon.append(min_epsilon + (max_epsilon - min_epsilon) * np.exp(-0.000001*i))", "_____no_output_____" ], [ "plt.plot(time, epsilon)\nplt.show()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import netCDF4 as nc\nimport numpy as np\nimport h5py", "_____no_output_____" ], [ "weight_file = '/export/home/srasp/tmp/015_1year_log_loss.h5'", "_____no_output_____" ], [ "f = h5py.File(weight_file, 'r')", "_____no_output_____" ], [ "layer_names = [n.decode('utf8') for n in f.attrs['layer_names']]; layer_names", "_____no_output_____" ], [ "g_tmp = f[layer_names[0]]; g_tmp", "_____no_output_____" ], [ "wn_tmp = [n.decode('utf8') for n in g_tmp.attrs['weight_names']]; wn_tmp", "_____no_output_____" ], [ "f.close()", "_____no_output_____" ], [ "def read_weights(weight_file):\n weights = []\n biases = []\n with h5py.File(weight_file, 'r') as f:\n layer_names = [n.decode('utf8') for n in f.attrs['layer_names']]\n for l in layer_names:\n g = f[l]\n weights.append(g[l+'/kernel:0'][:])\n biases.append(g[l+'/bias:0'][:])\n return weights, biases", "_____no_output_____" ], [ "w, b = read_weights(weight_file)", "_____no_output_____" ], [ "w[0].shape, b[0].shape", "_____no_output_____" ], [ "nchunk = 64", "_____no_output_____" ], [ "bre = b[0].reshape(1, -1)", "_____no_output_____" ], [ "np.savetxt('/export/home/srasp/tmp/test_bias.txt', bre, fmt='%.6e', delimiter=',')", "_____no_output_____" ], [ "wtrue = np.loadtxt('/export/home/srasp/tmp/results_015/layer1_kernel.txt', delimiter=',')", "_____no_output_____" ], [ "wtrue.shape", "_____no_output_____" ], [ "np.savetxt('/export/home/srasp/tmp/test_weight.txt', w[0].T, fmt='%.6e', delimiter=',')", "_____no_output_____" ], [ "# Convert norm and std files\nnorm_fn = '/scratch/srasp/preprocessed_data/purecrm_ess_train_sample1_norm.nc'", "_____no_output_____" ], [ "import xarray as xr", "_____no_output_____" ], [ "ds = nc.Dataset(norm_fn); ds", "_____no_output_____" ], [ "ds['feature_means'][:]", "_____no_output_____" ], [ "np.savetxt('/export/home/srasp/tmp/test_means.txt', ds['feature_means'][:].reshape(1, -1), fmt='%.6e', delimiter=',')", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# [fnmatch](https://docs.python.org/3/library/fnmatch.html)\n\n1. What is fnmatch and why is it useful?\n1. Why should I use fnmatch and not regex?\n1. Two examples\n\nFnmatch is part of the python standard library. Allows the use of UNIX style wildcards for string matching. Makes it easy to select a single file type out of a list (e.g. *.csv).\n\nWhile regex is much more powerful, fnmatch offers a simple syntax for using wildcards.\n\nIf you want to look for a string that starts with 5 characters, then a space and then 3 numbers between 4 and 7 you'll still need to resort to regex though.\n", "_____no_output_____" ], [ "## Simple example", "_____no_output_____" ] ], [ [ "import fnmatch\n\nFILES = [\"some_picture.png\", \"some_data.csv\", \"another_picture.png\"]\n\n# select only the .png files\nfor file in FILES:\n if fnmatch.fnmatch(file, '*.png'):\n print(file)\n \n# or using the fnmatch shorthand\nprint(fnmatch.filter(FILES, '*.png'))", "some_picture.png\nanother_picture.png\n['some_picture.png', 'another_picture.png']\n" ] ], [ [ "*SIDE NOTE*: The matching is **case insensitive**, if you want to perform a case sensitive match use [`fnmatch.fnmatchcase()`](https://docs.python.org/3/library/fnmatch.html#fnmatch.fnmatchcase)\n\n## Match a list of patterns\n", "_____no_output_____" ] ], [ [ "MODELS = [\"MESSAGEix-GLOBIOM 1.0\",\n \"MESSAGEix-GLOBIOM 1.1\",\n \"REMIND-MAgPIE 2.1-4.2\",\n \"REMIND-MAgPIE 1.7-3.2\",\n \"NIGEM\",\n \"POLES GECO2019\",\n \"COFFEE 1.0\",\n \"COFFEE 2.0\",\n \"TEA\",\n \"GCAM5.2\",\n \"GCAM5.3\"]\n\nMATCH_MODELS = [\"MESSAGEix-GLOBIOM*\", \"REMIND-MAgPIE*\"]\n\nmatch_any = lambda x, patterns: any(fnmatch.fnmatch(x, pattern) for pattern in patterns)\n\nfor m in MODELS:\n if match_any(m, MATCH_MODELS):\n print(m)", "MESSAGEix-GLOBIOM 1.0\nMESSAGEix-GLOBIOM 1.1\nREMIND-MAgPIE 2.1-4.2\nREMIND-MAgPIE 1.7-3.2\n" ] ] ]

[ "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "\nQuestion 6:Write a code in python to display different functions of python module.", "_____no_output_____" ] ], [ [ "#module required\nimport time\nprint(\"I am Iron Man.\")\ntime.sleep(2.4)#this function delays the time\nprint(\"I love you 3000.\") #this statement is printed after 2.4 seconds", "_____no_output_____" ], [ "\nimport time\n\n# seconds passed since epoch\nseconds = 1545925769.9618232\nlocal_time = time.ctime(seconds)\nprint(\"Local time:\", local_time)", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code" ] ]

[ "RSA-MD" ]

[ "RSA-MD" ]

[ "RSA-MD" ]

[ [ [ "# **3.d Formación de vectores**\n\n**Responsable:**\n\nCésar Zamora Martínez\n\n**Infraestructura usada:** \nGoogle Colab, para pruebas\n", "_____no_output_____" ], [ "## 0. Importamos librerias necesarias\n\n**Fuente:** 3c_formacion_matrices.ipynb, 3c_formacion_abc.ipynb, 3c_formacion_delta.ipynb", "_____no_output_____" ] ], [ [ "!curl https://colab.chainer.org/install | sh -", " % Total % Received % Xferd Average Speed Time Time Time Current\n Dload Upload Total Spent Left Speed\n100 1580 100 1580 0 0 3550 0 --:--:-- --:--:-- --:--:-- 3542\n+ apt -y -q install cuda-libraries-dev-10-0\nReading package lists...\nBuilding dependency tree...\nReading state information...\ncuda-libraries-dev-10-0 is already the newest version (10.0.130-1).\n0 upgraded, 0 newly installed, 0 to remove and 29 not upgraded.\n+ pip install -q cupy-cuda100 chainer \n\u001b[K |████████████████████████████████| 335.2MB 47kB/s \n\u001b[?25h+ set +ex\nInstallation succeeded!\n" ], [ "import cupy as cp\n\ndef formar_vectores(mu, Sigma):\n '''\n Calcula las cantidades u = \\Sigma^{-1} \\mu y v := \\Sigma^{-1} \\cdot 1 del problema de Markowitz\n\n Args:\n mu (cupy array, vector): valores medios esperados de activos (dimension n)\n Sigma (cupy array, matriz): matriz de covarianzas asociada a activos (dimension n x n)\n\n Return:\n u (cupy array, escalar): vector dado por \\cdot Sigma^-1 \\cdot mu (dimension n)\n v (cupy array, escalar): vector dado por Sigma^-1 \\cdot 1 (dimension n)\n '''\n\n # Vector auxiliar con entradas igual a 1\n n = Sigma.shape[0]\n ones_vector = cp.ones(n)\n\n # Formamos vector \\cdot Sigma^-1 mu y Sigm^-1 1\n # Nota: \n # 1) u= Sigma^-1 \\cdot mu se obtiene resolviendo Sigma u = mu\n # 2) v= Sigma^-1 \\cdot 1 se obtiene resolviendo Sigma v = 1\n\n # Obtiene vectores de interes\n u = cp.linalg.solve(Sigma, mu)\n u = u.transpose()[0] # correcion de expresion de array\n v = cp.linalg.solve(Sigma, ones_vector)\n\n return u , v", "_____no_output_____" ], [ "def formar_abc(mu, Sigma):\n '''\n Calcula las cantidades A, B y C del diagrama de flujo del problema de Markowitz\n\n Args:\n mu (cupy array, vector): valores medios esperados de activos (dimension n)\n Sigma (cupy array, matriz): matriz de covarianzas asociada a activos (dimension n x n)\n\n Return:\n A (cupy array, escalar): escalar dado por mu^t \\cdot Sigma^-1 \\cdot mu\n B (cupy array, escalar): escalar dado por 1^t \\cdot Sigma^-1 \\cdot 1\n C (cupy array, escalar): escalar dado por 1^t \\cdot Sigma^-1 \\cdot mu\n '''\n\n # Vector auxiliar con entradas igual a 1\n n = Sigma.shape[0]\n ones_vector = cp.ones(n)\n\n # Formamos vector \\cdot Sigma^-1 mu y Sigm^-1 1\n # Nota: \n # 1) u= Sigma^-1 \\cdot mu se obtiene resolviendo Sigma u = mu\n # 2) v= Sigma^-1 \\cdot 1 se obtiene resolviendo Sigma v = 1\n\n u, v = formar_vectores(mu, Sigma)\n\n # Obtiene escalares de interes\n A = mu.transpose()@u\n B = ones_vector.transpose()@v\n C = ones_vector.transpose()@u\n\n return A, B, C", "_____no_output_____" ], [ "def delta(A,B,C):\n '''\n Calcula las cantidad Delta = AB-C^2 del diagrama de flujo del problema de Markowitz\n\n Args:\n A (cupy array, escalar): escalar dado por mu^t \\cdot Sigma^-1 \\cdot mu\n B (cupy array, escalar): escalar dado por 1^t \\cdot Sigma^-1 \\cdot 1\n C (cupy array, escalar): escalar dado por 1^t \\cdot Sigma^-1 \\cdot mu\n\n Return:\n Delta (cupy array, escalar): escalar dado \\mu^t \\cdot \\Sigma^{-1} \\cdot \\mu\n '''\n Delta = A*B-C**2\n\n return Delta", "_____no_output_____" ] ], [ [ "## 1. Implementación\n\n**Consideraciones:**. Esta etapa supone que se conocen $\\bar{r}$, $\\mu$ y $\\Sigma$ asociados a los activos, ello con el objeto de es obtener valores escalares que serán relevantes para obtener los pesos del portafolio para el inversionista. Hasta este punto se asume que ya conocemos todos los términos presentes en las expresiones:\n\n$$A = \\mu^t \\cdot \\Sigma^{-1} \\cdot \\mu $$\n\n$$B = 1^t \\cdot \\Sigma^{-1} \\cdot 1 $$\n\n$$C = 1^t \\cdot \\Sigma^{-1} \\cdot \\mu = \\mu^t \\cdot \\Sigma^{-1} \\cdot 1 $$\n\nPara con ello poder estimar los multiplicadores de Lagrange asociados al problema:\n\n$$ w_0 = \\frac{1}{\\Delta} ( \\hat{r} \\cdot B - C ) $$\n\n$$ w_1 = \\frac{1}{\\Delta} (A - C \\cdot \\hat{r}) $$\n\nCon los que se forma la solución del sistema dada por\n\n$$w = w_0 \\cdot (\\Sigma^{-1} \\mu) + w_1 \\cdot (\\Sigma^{-1} 1) $$\n\nEn seguida se presenta el código correspondiente:", "_____no_output_____" ] ], [ [ "def formar_omegas(r, mu, Sigma):\n '''\n Calcula las cantidades w_o y w_1 del problema de Markowitz\n (valores de multiplicadores de Lagrange)\n\n Args:\n r (cupy array, escalar): escalar que denota el retorno esperado por el inversionista\n mu (cupy array, vector): valores medios esperados de activos (dimension n)\n Sigma (cupy array, matriz): matriz de covarianzas asociada a activos (dimension n x n)\n\n Return:\n w_0 (cupy array, escalar): escalar dada por \n w_0 = \\frac{1}{\\Delta} (B \\Sigma^{-1} \\hat{\\mu}- C\\Sigma^{-1} 1)\nw_1 (cupy array, escalar): escalar dado por \n w_1 = \\frac{1}{\\Delta} (C \\Sigma^{-1} \\hat{\\mu}- A\\Sigma^{-1} 1)\n '''\n # Obtenemos u = Sigma^{-1} \\hat{\\mu}, v = \\Sigma^{-1} 1\n u, v = formar_vectores(mu, Sigma)\n # Escalares relevantes\n A, B, C = formar_abc(mu, Sigma)\n Delta = delta(A,B,C)\n # Formamos w_0 y w_1\n w_0 = (1/Delta)*(r*B-C)\n w_1 = (1/Delta)*(A-C*r)\n\n return w_0, w_1", "_____no_output_____" ] ], [ [ "## 1.1 Valores de prueba", "_____no_output_____" ] ], [ [ "n= 10\n\n# r y mu\nr= 10\nmu=cp.random.rand(n, 1)\n\n# Sigma\nS=cp.random.rand(n, n)\nSigma=S@S\n\n# multiplicadores de lagrande\nformar_omegas(r,mu,Sigma)", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "run_checks = False\nrun_sample = False", "_____no_output_____" ] ], [ [ "### Overview\nThis notebook works on the IEEE-CIS Fraud Detection competition. Here I build a simple XGBoost model based on a balanced dataset.", "_____no_output_____" ], [ "### Lessons:\n\n. keep the categorical variables as single items\n\n. Use a high max_depth for xgboost (maybe 40)\n\n\n### Ideas to try:\n\n. train divergence of expected value (eg. for TransactionAmt and distance based on the non-fraud subset (not all subset as in the case now)\n\n. try using a temporal approach to CV", "_____no_output_____" ] ], [ [ "# all imports necessary for this notebook\n%matplotlib inline\nimport pandas as pd\nimport numpy as np\nimport matplotlib.pyplot as plt\nimport random\nimport gc\nimport copy\nimport missingno as msno \nimport xgboost\nfrom xgboost import XGBClassifier, XGBRegressor\nfrom sklearn.model_selection import StratifiedKFold, cross_validate, train_test_split \nfrom sklearn.metrics import roc_auc_score, r2_score\n\nimport warnings\nwarnings.filterwarnings('ignore')\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n for filename in filenames:\n print(os.path.join(dirname, filename))", "/kaggle/input/ieee-fraud-detection/test_identity.csv\n/kaggle/input/ieee-fraud-detection/sample_submission.csv\n/kaggle/input/ieee-fraud-detection/train_identity.csv\n/kaggle/input/ieee-fraud-detection/train_transaction.csv\n/kaggle/input/ieee-fraud-detection/test_transaction.csv\n/kaggle/input/ieee-preprocessed/master_df_top_all.csv\n/kaggle/input/ieee-preprocessed/master_df_top_300.csv\n/kaggle/input/ieee-preprocessed/master_df_top_100.csv\n/kaggle/input/ieee-preprocessed/master_df_top_200.csv\n" ], [ "# Helpers\n \ndef seed_everything(seed=0):\n '''Seed to make all processes deterministic '''\n random.seed(seed)\n os.environ['PYTHONHASHSEED'] = str(seed)\n np.random.seed(seed)\n \ndef drop_correlated_cols(df, threshold, sample_frac = 1):\n '''Drops one of two dataframe's columns whose pairwise pearson's correlation is above the provided threshold'''\n if sample_frac != 1:\n dataset = df.sample(frac = sample_frac).copy()\n else:\n dataset = df\n \n col_corr = set() # Set of all the names of deleted columns\n corr_matrix = dataset.corr()\n for i in range(len(corr_matrix.columns)):\n if corr_matrix.columns[i] in col_corr:\n continue\n for j in range(i):\n if (corr_matrix.iloc[i, j] >= threshold) and (corr_matrix.columns[j] not in col_corr):\n colname = corr_matrix.columns[i] # getting the name of column\n col_corr.add(colname)\n del dataset\n gc.collect()\n df.drop(columns = col_corr, inplace = True)\n\ndef calc_feature_difference(df, feature_name, indep_features, min_r2 = 0.1, min_r2_improv = 0, frac1 = 0.1, \n max_depth_start = 2, max_depth_step = 4):\n \n from copy import deepcopy\n \n print(\"Feature name %s\" %feature_name)\n #print(\"Indep_features %s\" %indep_features)\n \n is_imrpoving = True\n curr_max_depth = max_depth_start\n best_r2 = float(\"-inf\")\n clf_best = np.nan\n \n while is_imrpoving:\n clf = XGBRegressor(max_depth = curr_max_depth)\n\n rand_sample_indeces = df[df[feature_name].notnull()].sample(frac = frac1).index\n clf.fit(df.loc[rand_sample_indeces, indep_features], df.loc[rand_sample_indeces, feature_name]) \n\n rand_sample_indeces = df[df[feature_name].notnull()].sample(frac = frac1).index\n \n pred_y = clf.predict(df.loc[rand_sample_indeces, indep_features])\n r2Score = r2_score(df.loc[rand_sample_indeces, feature_name], pred_y)\n print(\"%d, R2 score %.4f\" % (curr_max_depth, r2Score))\n \n curr_max_depth = curr_max_depth + max_depth_step\n \n if r2Score > best_r2:\n best_r2 = r2Score\n clf_best = deepcopy(clf)\n if r2Score < best_r2 + (best_r2 * min_r2_improv) or (curr_max_depth > max_depth_start * max_depth_step and best_r2 < min_r2 / 2):\n is_imrpoving = False\n\n print(\"The best R2 score of %.4f\" % ( best_r2))\n \n if best_r2 > min_r2:\n pred_feature = clf_best.predict(df.loc[:, indep_features])\n return (df[feature_name] - pred_feature), best_r2\n else:\n return df[feature_name], best_r2", "_____no_output_____" ], [ "seed_everything()\npd.set_option('display.max_columns', 500)", "_____no_output_____" ], [ "master_df = pd.read_csv('/kaggle/input/ieee-preprocessed/master_df_top_300.csv')\nmaster_df.head()", "_____no_output_____" ], [ "cols_cat = {'id_12', 'id_13', 'id_14', 'id_15', 'id_16', 'id_17', 'id_18', 'id_19', 'id_20', 'id_21', 'id_22', \n 'id_23', 'id_24', 'id_25', 'id_26', 'id_27', 'id_28', 'id_29', 'id_30', 'id_31', 'id_32', 'id_33', \n 'id_34', 'id_35', 'id_36', 'id_37', 'id_38', 'DeviceType', 'DeviceInfo', 'ProductCD', 'card4', \n 'card6', 'M4','P_emaildomain', 'R_emaildomain', 'card1', 'card2', 'card3', 'card5', 'addr1', \n 'addr2', 'M1', 'M2', 'M3', 'M5', 'M6', 'M7', 'M8', 'M9'}", "_____no_output_____" ], [ "%%time\nindep_features = ['weekday', 'hours', 'TransactionDT', 'ProductCD', 'card1', 'card2', 'card3', 'card4', 'card5'\n , 'card6', 'addr1', 'addr2']\n\nfor feature in indep_features:\n master_df[feature] = master_df[feature].astype('category').cat.codes\n\ncont_cols_list = list(master_df.select_dtypes(include='number').columns)\ncont_features_list = [x for x in cont_cols_list if x not in cols_cat and x not in indep_features and x not in ['TransactionID', 'isFraud', 'TransactionDT', 'is_train_df']]\n\nfor cont_feature in cont_features_list:\n print(cont_feature)\n master_df[cont_feature], best_r2 = calc_feature_difference(master_df, cont_feature, indep_features, frac1= 0.025)\n if best_r2 > 0.9:\n master_df.drop(columns = [cont_feature], inplace = True)\n print(80 * '-')", "TransactionAmt\nFeature name TransactionAmt\n[07:15:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0920\n[07:15:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:15:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0798\nThe best R2 score of 0.0920\n--------------------------------------------------------------------------------\ndist1\nFeature name dist1\n[07:15:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0057\n[07:15:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:15:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0028\nThe best R2 score of 0.0057\n--------------------------------------------------------------------------------\ndist2\nFeature name dist2\n[07:15:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.1044\n[07:15:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:15:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0602\n[07:15:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.0602\n--------------------------------------------------------------------------------\nC1\nFeature name C1\n[07:15:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3002\n[07:15:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:15:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7461\n[07:15:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:15:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.5264\nThe best R2 score of 0.7461\n--------------------------------------------------------------------------------\nC2\nFeature name C2\n[07:16:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2321\n[07:16:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:16:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6294\n[07:16:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:16:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1139\nThe best R2 score of 0.6294\n--------------------------------------------------------------------------------\nC4\nFeature name C4\n[07:16:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3506\n[07:16:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:17:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4649\n[07:17:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:17:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4466\nThe best R2 score of 0.4649\n--------------------------------------------------------------------------------\nC5\nFeature name C5\n[07:17:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0255\n[07:17:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:17:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0192\nThe best R2 score of 0.0255\n--------------------------------------------------------------------------------\nC6\nFeature name C6\n[07:17:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2527\n[07:17:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:17:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2919\n[07:18:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:18:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.3616\n[07:18:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:18:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.4728\n[07:18:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:18:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n18, R2 score 0.3745\nThe best R2 score of 0.4728\n--------------------------------------------------------------------------------\nC8\nFeature name C8\n[07:19:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3817\n[07:19:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:19:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5930\n[07:19:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:19:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7405\n[07:19:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:19:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.8135\n[07:19:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:19:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n18, R2 score 0.9043\n[07:20:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:20:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n22, R2 score 0.7858\nThe best R2 score of 0.9043\n--------------------------------------------------------------------------------\nC9\nFeature name C9\n[07:20:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0286\n[07:20:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:20:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0256\nThe best R2 score of 0.0286\n--------------------------------------------------------------------------------\nC10\nFeature name C10\n[07:21:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4407\n[07:21:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:21:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8598\n[07:21:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:21:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7704\nThe best R2 score of 0.8598\n--------------------------------------------------------------------------------\nC11\nFeature name C11\n[07:21:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2542\n[07:21:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:21:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5071\n[07:22:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:22:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4696\nThe best R2 score of 0.5071\n--------------------------------------------------------------------------------\nC12\nFeature name C12\n[07:22:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3876\n[07:22:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:22:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5854\n[07:22:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:22:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4035\nThe best R2 score of 0.5854\n--------------------------------------------------------------------------------\nC13\nFeature name C13\n[07:23:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0894\n[07:23:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:23:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3133\n[07:23:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:23:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2588\nThe best R2 score of 0.3133\n--------------------------------------------------------------------------------\nC14\nFeature name C14\n[07:23:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1727\n[07:23:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:23:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3742\n[07:23:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:23:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.5662\n[07:24:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:24:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.3651\nThe best R2 score of 0.5662\n--------------------------------------------------------------------------------\nD1\nFeature name D1\n[07:24:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0848\n[07:24:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:24:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1039\n[07:24:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:24:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1027\nThe best R2 score of 0.1039\n--------------------------------------------------------------------------------\nD2\nFeature name D2\n[07:25:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0431\n[07:25:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:25:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0464\n[07:25:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0464\n--------------------------------------------------------------------------------\nD3\nFeature name D3\n[07:25:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0205\n[07:25:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:25:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0009\nThe best R2 score of 0.0205\n--------------------------------------------------------------------------------\nD4\nFeature name D4\n[07:25:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0740\n[07:25:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:25:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0817\n[07:25:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:25:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0556\nThe best R2 score of 0.0817\n--------------------------------------------------------------------------------\nD5\nFeature name D5\n[07:26:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0338\n[07:26:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:26:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0301\nThe best R2 score of 0.0338\n--------------------------------------------------------------------------------\nD6\nFeature name D6\n[07:26:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0849\n[07:26:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:26:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0704\nThe best R2 score of 0.0849\n--------------------------------------------------------------------------------\nD7\nFeature name D7\n[07:26:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0792\n[07:26:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:26:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0115\nThe best R2 score of 0.0792\n--------------------------------------------------------------------------------\nD8\nFeature name D8\n[07:26:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0755\n[07:26:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:26:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0534\nThe best R2 score of 0.0755\n--------------------------------------------------------------------------------\nD9\nFeature name D9\n[07:26:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[07:26:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:26:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[07:26:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:26:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nD10\nFeature name D10\n[07:26:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1127\n[07:26:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:26:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1124\nThe best R2 score of 0.1127\n--------------------------------------------------------------------------------\nD11\nFeature name D11\n[07:26:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0524\n[07:26:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:26:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0514\nThe best R2 score of 0.0524\n--------------------------------------------------------------------------------\nD12\nFeature name D12\n[07:26:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0199\n[07:26:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:27:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0358\nThe best R2 score of 0.0199\n--------------------------------------------------------------------------------\nD13\nFeature name D13\n[07:27:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0592\n[07:27:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:27:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0045\nThe best R2 score of 0.0592\n--------------------------------------------------------------------------------\nD14\nFeature name D14\n[07:27:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0457\n[07:27:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:27:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0309\nThe best R2 score of 0.0457\n--------------------------------------------------------------------------------\nD15\nFeature name D15\n[07:27:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1118\n[07:27:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:27:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1179\n[07:27:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:27:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0978\nThe best R2 score of 0.1179\n--------------------------------------------------------------------------------\nV2\nFeature name V2\n[07:27:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0002\n[07:27:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:27:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0255\nThe best R2 score of -0.0002\n--------------------------------------------------------------------------------\nV3\nFeature name V3\n[07:27:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0024\n[07:27:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:27:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0102\nThe best R2 score of 0.0024\n--------------------------------------------------------------------------------\nV4\nFeature name V4\n[07:28:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0036\n[07:28:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:28:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0001\nThe best R2 score of 0.0036\n--------------------------------------------------------------------------------\nV5\nFeature name V5\n[07:28:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0029\n[07:28:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:28:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0100\n[07:28:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0100\n--------------------------------------------------------------------------------\nV6\nFeature name V6\n[07:28:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0054\n[07:28:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:28:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0218\nThe best R2 score of 0.0054\n--------------------------------------------------------------------------------\nV7\nFeature name V7\n[07:28:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0003\n[07:28:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:28:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0052\nThe best R2 score of 0.0003\n--------------------------------------------------------------------------------\nV10\nFeature name V10\n[07:28:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0566\n[07:28:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:28:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0730\n[07:28:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:28:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0412\nThe best R2 score of 0.0730\n--------------------------------------------------------------------------------\nV12\nFeature name V12\n[07:29:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3424\n[07:29:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:29:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3361\nThe best R2 score of 0.3424\n--------------------------------------------------------------------------------\nV13\nFeature name V13\n[07:29:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3457\n[07:29:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:29:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3468\n[07:29:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:29:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.3314\nThe best R2 score of 0.3468\n--------------------------------------------------------------------------------\nV19\nFeature name V19\n[07:29:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0933\n[07:30:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:30:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1043\n[07:30:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:30:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0796\nThe best R2 score of 0.1043\n--------------------------------------------------------------------------------\nV20\nFeature name V20\n[07:30:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0804\n[07:30:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:30:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0888\n[07:30:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:30:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0809\nThe best R2 score of 0.0888\n--------------------------------------------------------------------------------\nV21\nFeature name V21\n[07:31:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8833\n[07:31:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:31:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8940\n[07:31:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:31:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.8853\nThe best R2 score of 0.8940\n--------------------------------------------------------------------------------\nV23\nFeature name V23\n[07:31:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0191\n[07:31:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:31:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1531\nThe best R2 score of 0.0191\n--------------------------------------------------------------------------------\nV24\nFeature name V24\n[07:31:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0169\n[07:31:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:32:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0180\n[07:32:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0180\n--------------------------------------------------------------------------------\nV25\nFeature name V25\n[07:32:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0913\n[07:32:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:32:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2073\n[07:32:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:32:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0292\nThe best R2 score of 0.2073\n--------------------------------------------------------------------------------\nV29\nFeature name V29\n[07:32:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1793\n[07:32:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:32:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1957\n[07:33:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:33:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1852\nThe best R2 score of 0.1957\n--------------------------------------------------------------------------------\nV34\nFeature name V34\n[07:33:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7948\n[07:33:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:33:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7568\nThe best R2 score of 0.7948\n--------------------------------------------------------------------------------\nV35\nFeature name V35\n[07:33:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3556\n[07:33:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:33:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3482\nThe best R2 score of 0.3556\n--------------------------------------------------------------------------------\nV36\nFeature name V36\n[07:34:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3595\n[07:34:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:34:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3517\nThe best R2 score of 0.3595\n--------------------------------------------------------------------------------\nV37\nFeature name V37\n[07:34:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0326\n[07:34:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:34:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0209\nThe best R2 score of 0.0326\n--------------------------------------------------------------------------------\nV38\nFeature name V38\n[07:34:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0804\n[07:34:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:34:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0800\nThe best R2 score of 0.0804\n--------------------------------------------------------------------------------\nV39\nFeature name V39\n[07:34:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7541\n[07:34:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:34:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7180\nThe best R2 score of 0.7541\n--------------------------------------------------------------------------------\nV40\nFeature name V40\n[07:35:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6474\n[07:35:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:35:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6398\nThe best R2 score of 0.6474\n--------------------------------------------------------------------------------\nV44\nFeature name V44\n[07:35:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0549\n[07:35:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:35:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2229\n[07:35:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:35:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.0446\nThe best R2 score of 0.2229\n--------------------------------------------------------------------------------\nV45\nFeature name V45\n[07:35:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0066\n[07:36:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:36:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0077\n[07:36:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0077\n--------------------------------------------------------------------------------\nV47\nFeature name V47\n[07:36:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0234\n[07:36:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:36:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0081\nThe best R2 score of 0.0234\n--------------------------------------------------------------------------------\nV48\nFeature name V48\n[07:36:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1936\n[07:36:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:36:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2121\n[07:36:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:36:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2018\nThe best R2 score of 0.2121\n--------------------------------------------------------------------------------\nV49\nFeature name V49\n[07:36:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1844\n[07:37:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:37:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1963\n[07:37:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:37:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1855\nThe best R2 score of 0.1963\n--------------------------------------------------------------------------------\nV51\nFeature name V51\n[07:37:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8429\n[07:37:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:37:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8608\n[07:37:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:37:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.8385\nThe best R2 score of 0.8608\n--------------------------------------------------------------------------------\nV52\nFeature name V52\n[07:38:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7573\n[07:38:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:38:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7688\n[07:38:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:38:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7489\nThe best R2 score of 0.7688\n--------------------------------------------------------------------------------\nV53\nFeature name V53\n[07:38:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3724\n[07:38:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:38:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3769\n[07:38:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:38:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.3607\nThe best R2 score of 0.3769\n--------------------------------------------------------------------------------\nV54\nFeature name V54\n[07:39:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3852\n[07:39:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:39:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3797\nThe best R2 score of 0.3852\n--------------------------------------------------------------------------------\nV55\nFeature name V55\n[07:39:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0330\n[07:39:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:39:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0053\nThe best R2 score of 0.0330\n--------------------------------------------------------------------------------\nV56\nFeature name V56\n[07:39:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0610\n[07:39:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:39:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1407\n[07:40:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:40:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0646\nThe best R2 score of 0.1407\n--------------------------------------------------------------------------------\nV60\nFeature name V60\n[07:40:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7122\n[07:40:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:40:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7488\n[07:40:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:40:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7173\nThe best R2 score of 0.7488\n--------------------------------------------------------------------------------\nV61\nFeature name V61\n[07:40:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0887\n[07:41:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:41:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1046\n[07:41:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:41:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1016\nThe best R2 score of 0.1046\n--------------------------------------------------------------------------------\nV62\nFeature name V62\n[07:41:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0630\n[07:41:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:41:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0710\n[07:41:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:41:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0888\n[07:42:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:42:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.0529\nThe best R2 score of 0.0888\n--------------------------------------------------------------------------------\nV64\nFeature name V64\n[07:42:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7319\n[07:42:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:42:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7874\n[07:42:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:42:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7459\nThe best R2 score of 0.7874\n--------------------------------------------------------------------------------\nV66\nFeature name V66\n[07:42:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0552\n[07:42:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:43:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0800\n[07:43:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:43:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0136\nThe best R2 score of 0.0800\n--------------------------------------------------------------------------------\nV67\nFeature name V67\n[07:43:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0476\n[07:43:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:43:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0492\n[07:43:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0492\n--------------------------------------------------------------------------------\nV69\nFeature name V69\n[07:43:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1823\n[07:43:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:43:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1969\n[07:43:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:43:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1905\nThe best R2 score of 0.1969\n--------------------------------------------------------------------------------\nV70\nFeature name V70\n[07:44:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1644\n[07:44:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:44:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1962\n[07:44:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:44:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1762\nThe best R2 score of 0.1962\n--------------------------------------------------------------------------------\nV74\nFeature name V74\n[07:44:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7409\n[07:44:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:45:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7244\nThe best R2 score of 0.7409\n--------------------------------------------------------------------------------\nV75\nFeature name V75\n[07:45:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3092\n[07:45:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:45:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3103\n[07:45:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:45:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2904\nThe best R2 score of 0.3103\n--------------------------------------------------------------------------------\nV76\nFeature name V76\n[07:45:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3192\n[07:45:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:45:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3164\nThe best R2 score of 0.3192\n--------------------------------------------------------------------------------\nV77\nFeature name V77\n[07:46:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0677\n[07:46:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:46:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1637\n[07:46:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:46:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.6370\nThe best R2 score of 0.1637\n--------------------------------------------------------------------------------\nV78\nFeature name V78\n[07:46:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0651\n[07:46:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:46:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2534\n[07:47:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:47:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.3201\nThe best R2 score of 0.2534\n--------------------------------------------------------------------------------\nV79\nFeature name V79\n[07:47:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8942\n[07:47:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:47:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8795\nThe best R2 score of 0.8942\n--------------------------------------------------------------------------------\nV81\nFeature name V81\n[07:47:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7104\n[07:47:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:47:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6667\nThe best R2 score of 0.7104\n--------------------------------------------------------------------------------\nV82\nFeature name V82\n[07:47:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0991\n[07:48:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:48:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1028\n[07:48:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:48:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0879\nThe best R2 score of 0.1028\n--------------------------------------------------------------------------------\nV83\nFeature name V83\n[07:48:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0780\n[07:48:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:48:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0775\nThe best R2 score of 0.0780\n--------------------------------------------------------------------------------\nV86\nFeature name V86\n[07:48:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0114\n[07:48:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:49:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0485\n[07:49:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0485\n--------------------------------------------------------------------------------\nV87\nFeature name V87\n[07:49:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0294\n[07:49:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:49:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2333\n[07:49:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:49:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.7488\nThe best R2 score of 0.2333\n--------------------------------------------------------------------------------\nV90\nFeature name V90\n[07:49:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1835\n[07:49:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:49:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2071\n[07:50:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:50:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1868\nThe best R2 score of 0.2071\n--------------------------------------------------------------------------------\nV91\nFeature name V91\n[07:50:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1733\n[07:50:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:50:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1938\n[07:50:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:50:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1767\nThe best R2 score of 0.1938\n--------------------------------------------------------------------------------\nV94\nFeature name V94\n[07:50:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9111\n[07:51:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:51:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9116\n[07:51:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:51:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9116\n[07:51:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:51:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9061\nThe best R2 score of 0.9116\n--------------------------------------------------------------------------------\nV95\nFeature name V95\n[07:51:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6377\n[07:51:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:52:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9837\n[07:52:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:52:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9315\nThe best R2 score of 0.9837\n--------------------------------------------------------------------------------\nV96\nFeature name V96\n[07:52:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6587\n[07:52:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:52:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9740\n[07:52:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:52:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9798\n[07:53:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:53:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9776\nThe best R2 score of 0.9798\n--------------------------------------------------------------------------------\nV97\nFeature name V97\n[07:53:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6661\n[07:53:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:53:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9909\n[07:53:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:53:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9785\nThe best R2 score of 0.9909\n--------------------------------------------------------------------------------\nV99\nFeature name V99\n[07:54:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1171\n[07:54:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:54:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2309\n[07:54:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:54:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2118\nThe best R2 score of 0.2309\n--------------------------------------------------------------------------------\nV100\nFeature name V100\n[07:55:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1341\n[07:55:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:55:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2255\n[07:55:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:55:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1572\nThe best R2 score of 0.2255\n--------------------------------------------------------------------------------\nV102\nFeature name V102\n[07:55:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5246\n[07:55:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:55:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8036\n[07:55:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:56:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9270\n[07:56:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:56:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9832\n[07:56:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:56:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n18, R2 score 0.8002\nThe best R2 score of 0.9832\n--------------------------------------------------------------------------------\nV103\nFeature name V103\n[07:57:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6407\n[07:57:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:57:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9855\n[07:57:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:57:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.8247\nThe best R2 score of 0.9855\n--------------------------------------------------------------------------------\nV104\nFeature name V104\n[07:57:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4281\n[07:58:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:58:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3100\nThe best R2 score of 0.4281\n--------------------------------------------------------------------------------\nV105\nFeature name V105\n[07:58:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7554\n[07:58:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:58:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8688\n[07:58:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:58:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.8674\nThe best R2 score of 0.8688\n--------------------------------------------------------------------------------\nV109\nFeature name V109\n[07:59:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0047\n[07:59:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:59:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0138\nThe best R2 score of 0.0047\n--------------------------------------------------------------------------------\nV110\nFeature name V110\n[07:59:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0029\n[07:59:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:59:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0529\nThe best R2 score of 0.0029\n--------------------------------------------------------------------------------\nV112\nFeature name V112\n[07:59:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0010\n[07:59:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[07:59:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0008\nThe best R2 score of 0.0010\n--------------------------------------------------------------------------------\nV115\nFeature name V115\n[07:59:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0096\n[08:00:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:00:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0116\n[08:00:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0116\n--------------------------------------------------------------------------------\nV116\nFeature name V116\n[08:00:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0041\n[08:00:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:00:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0166\nThe best R2 score of 0.0041\n--------------------------------------------------------------------------------\nV117\nFeature name V117\n[08:00:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0414\n[08:00:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:00:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1067\nThe best R2 score of 0.0414\n--------------------------------------------------------------------------------\nV122\nFeature name V122\n[08:00:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0505\n[08:01:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:01:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0239\n[08:01:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.0239\n--------------------------------------------------------------------------------\nV124\nFeature name V124\n[08:01:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0587\n[08:01:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:01:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0483\nThe best R2 score of 0.0587\n--------------------------------------------------------------------------------\nV126\nFeature name V126\n[08:01:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5813\n[08:01:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:01:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2682\nThe best R2 score of 0.5813\n--------------------------------------------------------------------------------\nV127\nFeature name V127\n[08:01:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7013\n[08:02:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:02:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5785\nThe best R2 score of 0.7013\n--------------------------------------------------------------------------------\nV128\nFeature name V128\n[08:02:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4943\n[08:02:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:02:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5320\n[08:02:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:02:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7111\n[08:02:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:02:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score -1.0119\nThe best R2 score of 0.7111\n--------------------------------------------------------------------------------\nV129\nFeature name V129\n[08:03:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0111\n[08:03:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:03:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0114\nThe best R2 score of -0.0111\n--------------------------------------------------------------------------------\nV130\nFeature name V130\n[08:03:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0210\n[08:03:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:03:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0462\nThe best R2 score of 0.0210\n--------------------------------------------------------------------------------\nV131\nFeature name V131\n[08:04:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0185\n[08:04:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:04:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0197\nThe best R2 score of 0.0185\n--------------------------------------------------------------------------------\nV132\nFeature name V132\n[08:04:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6261\n[08:04:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:04:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9612\n[08:04:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:04:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.3090\nThe best R2 score of 0.9612\n--------------------------------------------------------------------------------\nV133\nFeature name V133\n[08:05:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5530\n[08:05:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:05:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8278\n[08:05:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:05:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9529\n[08:05:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:05:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.2752\nThe best R2 score of 0.9529\n--------------------------------------------------------------------------------\nV134\nFeature name V134\n[08:06:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4508\n[08:06:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:06:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9812\n[08:06:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:06:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.1467\nThe best R2 score of 0.9812\n--------------------------------------------------------------------------------\nV135\nFeature name V135\n[08:06:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.1941\n[08:06:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:06:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0019\n[08:07:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.0019\n--------------------------------------------------------------------------------\nV136\nFeature name V136\n[08:07:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -14.7716\n[08:07:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:07:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1319\n[08:07:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:07:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.3096\nThe best R2 score of 0.1319\n--------------------------------------------------------------------------------\nV137\nFeature name V137\n[08:07:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1638\n[08:07:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:07:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -1.2927\nThe best R2 score of 0.1638\n--------------------------------------------------------------------------------\nV139\nFeature name V139\n[08:08:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0415\n[08:08:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0432\n[08:08:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0432\n--------------------------------------------------------------------------------\nV141\nFeature name V141\n[08:08:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0394\n[08:08:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0087\nThe best R2 score of 0.0394\n--------------------------------------------------------------------------------\nV143\nFeature name V143\n[08:08:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7714\n[08:08:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9480\n[08:08:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7329\nThe best R2 score of 0.9480\n--------------------------------------------------------------------------------\nV144\nFeature name V144\n[08:08:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4921\n[08:08:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5504\n[08:08:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4726\nThe best R2 score of 0.5504\n--------------------------------------------------------------------------------\nV145\nFeature name V145\n[08:08:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4794\n[08:08:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4940\n[08:08:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.3922\nThe best R2 score of 0.4940\n--------------------------------------------------------------------------------\nV146\nFeature name V146\n[08:08:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0117\n[08:08:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0377\n[08:08:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0377\n--------------------------------------------------------------------------------\nV149\nFeature name V149\n[08:08:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0958\n[08:08:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1098\n[08:08:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:08:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.4213\nThe best R2 score of 0.1098\n--------------------------------------------------------------------------------\nV150\nFeature name V150\n[08:09:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4886\n[08:09:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4683\nThe best R2 score of 0.4886\n--------------------------------------------------------------------------------\nV151\nFeature name V151\n[08:09:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5041\n[08:09:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4977\nThe best R2 score of 0.5041\n--------------------------------------------------------------------------------\nV152\nFeature name V152\n[08:09:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4205\n[08:09:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4091\nThe best R2 score of 0.4205\n--------------------------------------------------------------------------------\nV154\nFeature name V154\n[08:09:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1359\n[08:09:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1235\nThe best R2 score of 0.1359\n--------------------------------------------------------------------------------\nV160\nFeature name V160\n[08:09:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3318\n[08:09:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3993\n[08:09:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2804\nThe best R2 score of 0.3993\n--------------------------------------------------------------------------------\nV161\nFeature name V161\n[08:09:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.9572\n[08:09:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.4251\n[08:09:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.4251\n--------------------------------------------------------------------------------\nV162\nFeature name V162\n[08:09:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.1922\n[08:09:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.3519\nThe best R2 score of -0.1922\n--------------------------------------------------------------------------------\nV164\nFeature name V164\n[08:09:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2818\n[08:09:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3343\n[08:09:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.6244\n[08:09:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.8383\n[08:09:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:09:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n18, R2 score 0.5811\nThe best R2 score of 0.8383\n--------------------------------------------------------------------------------\nV165\nFeature name V165\n[08:10:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3743\n[08:10:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:10:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4322\n[08:10:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:10:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4774\n[08:10:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:10:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.3527\nThe best R2 score of 0.4774\n--------------------------------------------------------------------------------\nV166\nFeature name V166\n[08:10:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4557\n[08:10:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:10:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5915\n[08:10:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:10:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.3719\nThe best R2 score of 0.5915\n--------------------------------------------------------------------------------\nV170\nFeature name V170\n[08:10:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0372\n[08:10:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:10:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0058\nThe best R2 score of 0.0372\n--------------------------------------------------------------------------------\nV171\nFeature name V171\n[08:10:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0757\n[08:10:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:10:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1169\n[08:11:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:11:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.0741\nThe best R2 score of 0.1169\n--------------------------------------------------------------------------------\nV173\nFeature name V173\n[08:11:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0179\n[08:11:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:11:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0323\nThe best R2 score of 0.0179\n--------------------------------------------------------------------------------\nV176\nFeature name V176\n[08:11:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2555\n[08:11:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:11:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5801\n[08:11:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:11:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.3614\nThe best R2 score of 0.5801\n--------------------------------------------------------------------------------\nV177\nFeature name V177\n[08:11:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5450\n[08:11:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:11:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7905\n[08:11:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:11:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9822\n[08:11:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:11:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.8925\nThe best R2 score of 0.9822\n--------------------------------------------------------------------------------\nV187\nFeature name V187\n[08:12:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0967\n[08:12:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1936\nThe best R2 score of -0.0967\n--------------------------------------------------------------------------------\nV188\nFeature name V188\n[08:12:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0115\n[08:12:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0267\n[08:12:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0267\n--------------------------------------------------------------------------------\nV189\nFeature name V189\n[08:12:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0003\n[08:12:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0410\nThe best R2 score of -0.0003\n--------------------------------------------------------------------------------\nV191\nFeature name V191\n[08:12:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0122\n[08:12:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.2211\nThe best R2 score of -0.0122\n--------------------------------------------------------------------------------\nV192\nFeature name V192\n[08:12:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0425\n[08:12:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0036\n[08:12:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0036\n--------------------------------------------------------------------------------\nV194\nFeature name V194\n[08:12:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0388\n[08:12:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0031\nThe best R2 score of 0.0388\n--------------------------------------------------------------------------------\nV201\nFeature name V201\n[08:12:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0140\n[08:12:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0376\nThe best R2 score of 0.0140\n--------------------------------------------------------------------------------\nV202\nFeature name V202\n[08:12:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1315\n[08:12:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -7.4106\nThe best R2 score of 0.1315\n--------------------------------------------------------------------------------\nV203\nFeature name V203\n[08:12:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0924\n[08:12:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4865\n[08:12:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:12:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.0054\nThe best R2 score of 0.4865\n--------------------------------------------------------------------------------\nV206\nFeature name V206\n[08:12:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -1.0298\n[08:13:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1694\n[08:13:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.1694\n--------------------------------------------------------------------------------\nV208\nFeature name V208\n[08:13:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0072\n[08:13:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.9807\nThe best R2 score of -0.0072\n--------------------------------------------------------------------------------\nV209\nFeature name V209\n[08:13:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0330\n[08:13:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0011\n[08:13:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0011\n--------------------------------------------------------------------------------\nV211\nFeature name V211\n[08:13:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -2.4491\n[08:13:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1501\n[08:13:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.5734\n[08:13:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.0946\nThe best R2 score of 0.5734\n--------------------------------------------------------------------------------\nV214\nFeature name V214\n[08:13:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -40.1710\n[08:13:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3079\n[08:13:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.0709\nThe best R2 score of 0.3079\n--------------------------------------------------------------------------------\nV218\nFeature name V218\n[08:13:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2363\n[08:13:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4939\n[08:13:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:13:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4540\nThe best R2 score of 0.4939\n--------------------------------------------------------------------------------\nV221\nFeature name V221\n[08:14:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0075\n[08:14:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.3627\nThe best R2 score of 0.0075\n--------------------------------------------------------------------------------\nV222\nFeature name V222\n[08:14:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3026\n[08:14:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3859\n[08:14:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1023\nThe best R2 score of 0.3859\n--------------------------------------------------------------------------------\nV223\nFeature name V223\n[08:14:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0291\n[08:14:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0613\nThe best R2 score of 0.0291\n--------------------------------------------------------------------------------\nV224\nFeature name V224\n[08:14:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.3549\n[08:14:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0206\n[08:14:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0206\n--------------------------------------------------------------------------------\nV225\nFeature name V225\n[08:14:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.8025\n[08:14:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.5157\n[08:14:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.5157\n--------------------------------------------------------------------------------\nV226\nFeature name V226\n[08:14:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3572\n[08:14:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0067\nThe best R2 score of 0.3572\n--------------------------------------------------------------------------------\nV229\nFeature name V229\n[08:14:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0456\n[08:14:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0454\nThe best R2 score of 0.0456\n--------------------------------------------------------------------------------\nV230\nFeature name V230\n[08:14:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0829\n[08:14:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0917\nThe best R2 score of 0.0829\n--------------------------------------------------------------------------------\nV234\nFeature name V234\n[08:14:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1841\n[08:14:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5568\n[08:14:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:14:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.5633\n[08:15:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:15:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.2440\nThe best R2 score of 0.5633\n--------------------------------------------------------------------------------\nV236\nFeature name V236\n[08:15:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4358\n[08:15:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:15:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2918\nThe best R2 score of 0.4358\n--------------------------------------------------------------------------------\nV243\nFeature name V243\n[08:15:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.1768\n[08:15:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:15:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.4186\nThe best R2 score of -0.1768\n--------------------------------------------------------------------------------\nV244\nFeature name V244\n[08:15:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0129\n[08:15:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:15:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0054\nThe best R2 score of 0.0129\n--------------------------------------------------------------------------------\nV245\nFeature name V245\n[08:15:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0546\n[08:15:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:15:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0063\n[08:15:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.0063\n--------------------------------------------------------------------------------\nV247\nFeature name V247\n[08:15:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.1791\n[08:15:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:15:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0581\n[08:15:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.0581\n--------------------------------------------------------------------------------\nV248\nFeature name V248\n[08:15:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0855\n[08:15:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:15:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -1.0362\nThe best R2 score of 0.0855\n--------------------------------------------------------------------------------\nV249\nFeature name V249\n[08:15:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0032\n[08:15:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:15:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.7305\nThe best R2 score of -0.0032\n--------------------------------------------------------------------------------\nV253\nFeature name V253\n[08:15:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0012\n[08:15:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:15:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.7159\nThe best R2 score of -0.0012\n--------------------------------------------------------------------------------\nV254\nFeature name V254\n[08:15:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.5027\n[08:15:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.5741\nThe best R2 score of -0.5027\n--------------------------------------------------------------------------------\nV257\nFeature name V257\n[08:16:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3238\n[08:16:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5942\n[08:16:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.2051\nThe best R2 score of 0.5942\n--------------------------------------------------------------------------------\nV258\nFeature name V258\n[08:16:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0321\n[08:16:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0781\n[08:16:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.6362\nThe best R2 score of 0.0781\n--------------------------------------------------------------------------------\nV263\nFeature name V263\n[08:16:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1177\n[08:16:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0415\nThe best R2 score of 0.1177\n--------------------------------------------------------------------------------\nV264\nFeature name V264\n[08:16:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -7.5730\n[08:16:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1877\n[08:16:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.1877\n--------------------------------------------------------------------------------\nV265\nFeature name V265\n[08:16:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -5.1975\n[08:16:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.3002\n[08:16:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of -0.3002\n--------------------------------------------------------------------------------\nV266\nFeature name V266\n[08:16:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0222\n[08:16:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5373\n[08:16:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -8.9798\nThe best R2 score of 0.5373\n--------------------------------------------------------------------------------\nV267\nFeature name V267\n[08:16:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.1919\n[08:16:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:16:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -4.2503\nThe best R2 score of -0.1919\n--------------------------------------------------------------------------------\nV269\nFeature name V269\n[08:16:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1138\n[08:16:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:17:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.4789\nThe best R2 score of 0.1138\n--------------------------------------------------------------------------------\nV270\nFeature name V270\n[08:17:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0056\n[08:17:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:17:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0458\nThe best R2 score of -0.0056\n--------------------------------------------------------------------------------\nV271\nFeature name V271\n[08:17:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0914\n[08:17:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:17:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -1.1909\nThe best R2 score of -0.0914\n--------------------------------------------------------------------------------\nV272\nFeature name V272\n[08:17:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0023\n[08:17:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:17:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0222\n[08:17:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0222\n--------------------------------------------------------------------------------\nV273\nFeature name V273\n[08:17:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0045\n[08:17:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:17:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -1.2337\nThe best R2 score of 0.0045\n--------------------------------------------------------------------------------\nV279\nFeature name V279\n[08:17:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7365\n[08:17:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:17:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7703\n[08:17:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:17:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9881\n[08:17:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:18:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9437\nThe best R2 score of 0.9881\n--------------------------------------------------------------------------------\nV280\nFeature name V280\n[08:18:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6736\n[08:18:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:18:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9822\n[08:18:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:18:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.8636\nThe best R2 score of 0.9822\n--------------------------------------------------------------------------------\nV281\nFeature name V281\n[08:19:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0159\n[08:19:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:19:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1086\n[08:19:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:19:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.0895\nThe best R2 score of 0.1086\n--------------------------------------------------------------------------------\nV282\nFeature name V282\n[08:19:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0720\n[08:20:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:20:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0814\n[08:20:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:20:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0764\nThe best R2 score of 0.0814\n--------------------------------------------------------------------------------\nV283\nFeature name V283\n[08:20:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0785\n[08:20:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:20:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1996\n[08:20:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:20:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1299\nThe best R2 score of 0.1996\n--------------------------------------------------------------------------------\nV285\nFeature name V285\n[08:21:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0924\n[08:21:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:21:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2049\n[08:21:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:21:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2187\n[08:21:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:21:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.0817\nThe best R2 score of 0.2187\n--------------------------------------------------------------------------------\nV286\nFeature name V286\n[08:22:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0094\n[08:22:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:22:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0289\nThe best R2 score of 0.0094\n--------------------------------------------------------------------------------\nV287\nFeature name V287\n[08:22:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0675\n[08:22:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:22:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1704\n[08:22:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:22:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1054\nThe best R2 score of 0.1704\n--------------------------------------------------------------------------------\nV288\nFeature name V288\n[08:23:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0219\n[08:23:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:23:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0301\n[08:23:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0301\n--------------------------------------------------------------------------------\nV290\nFeature name V290\n[08:23:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1179\n[08:23:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:23:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2683\n[08:23:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:23:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0731\nThe best R2 score of 0.2683\n--------------------------------------------------------------------------------\nV291\nFeature name V291\n[08:24:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3416\n[08:24:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:24:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6289\n[08:24:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:24:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.6959\n[08:24:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:24:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.7846\n[08:24:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:24:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n18, R2 score 0.6818\nThe best R2 score of 0.7846\n--------------------------------------------------------------------------------\nV292\nFeature name V292\n[08:25:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4965\n[08:25:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:25:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8694\n[08:25:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:25:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.6871\nThe best R2 score of 0.8694\n--------------------------------------------------------------------------------\nV293\nFeature name V293\n[08:26:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5416\n[08:26:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:26:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9849\n[08:26:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:26:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9100\nThe best R2 score of 0.9849\n--------------------------------------------------------------------------------\nV294\nFeature name V294\n[08:26:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6528\n[08:27:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:27:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9870\n[08:27:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:27:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9673\nThe best R2 score of 0.9870\n--------------------------------------------------------------------------------\nV295\nFeature name V295\n[08:27:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6777\n[08:27:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:27:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9739\n[08:27:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:27:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9878\n[08:28:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:28:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9764\nThe best R2 score of 0.9878\n--------------------------------------------------------------------------------\nV296\nFeature name V296\n[08:28:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6256\n[08:28:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:28:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7943\n[08:28:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:29:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7844\nThe best R2 score of 0.7943\n--------------------------------------------------------------------------------\nV298\nFeature name V298\n[08:29:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6609\n[08:29:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:29:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7495\n[08:29:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:29:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.8282\n[08:29:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:29:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.8503\n[08:30:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:30:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n18, R2 score 0.7274\nThe best R2 score of 0.8503\n--------------------------------------------------------------------------------\nV301\nFeature name V301\n[08:30:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0015\n[08:30:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:31:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0200\nThe best R2 score of 0.0015\n--------------------------------------------------------------------------------\nV302\nFeature name V302\n[08:31:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8012\n[08:31:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:31:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7998\nThe best R2 score of 0.8012\n--------------------------------------------------------------------------------\nV303\nFeature name V303\n[08:31:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6599\n[08:31:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:31:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6962\n[08:31:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:31:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.6281\nThe best R2 score of 0.6962\n--------------------------------------------------------------------------------\nV304\nFeature name V304\n[08:32:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7264\n[08:32:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:32:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7472\n[08:32:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:32:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7254\nThe best R2 score of 0.7472\n--------------------------------------------------------------------------------\nV306\nFeature name V306\n[08:32:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2834\n[08:32:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:32:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1496\nThe best R2 score of 0.2834\n--------------------------------------------------------------------------------\nV307\nFeature name V307\n[08:33:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4737\n[08:33:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:33:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6765\n[08:33:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:33:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.6130\nThe best R2 score of 0.6765\n--------------------------------------------------------------------------------\nV308\nFeature name V308\n[08:33:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0127\n[08:33:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:33:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8979\n[08:33:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:34:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -1.2454\nThe best R2 score of 0.8979\n--------------------------------------------------------------------------------\nV309\nFeature name V309\n[08:34:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0145\n[08:34:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:34:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1178\nThe best R2 score of 0.0145\n--------------------------------------------------------------------------------\nV310\nFeature name V310\n[08:34:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0244\n[08:34:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:34:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.5758\nThe best R2 score of 0.0244\n--------------------------------------------------------------------------------\nV311\nFeature name V311\n[08:35:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0051\n[08:35:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:35:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0244\nThe best R2 score of -0.0051\n--------------------------------------------------------------------------------\nV312\nFeature name V312\n[08:35:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0049\n[08:35:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:35:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1351\nThe best R2 score of 0.0049\n--------------------------------------------------------------------------------\nV313\nFeature name V313\n[08:35:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0008\n[08:35:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:35:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0141\nThe best R2 score of -0.0008\n--------------------------------------------------------------------------------\nV314\nFeature name V314\n[08:35:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0203\n[08:36:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:36:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0870\n[08:36:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:36:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0842\nThe best R2 score of 0.0870\n--------------------------------------------------------------------------------\nV315\nFeature name V315\n[08:36:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0303\n[08:36:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:36:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0306\nThe best R2 score of -0.0303\n--------------------------------------------------------------------------------\nV316\nFeature name V316\n[08:36:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1164\n[08:36:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:36:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7611\n[08:37:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:37:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9761\n[08:37:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:37:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.6066\nThe best R2 score of 0.9761\n--------------------------------------------------------------------------------\nV317\nFeature name V317\n[08:37:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.4905\n[08:37:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:37:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8781\n[08:38:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:38:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4361\nThe best R2 score of 0.8781\n--------------------------------------------------------------------------------\nV318\nFeature name V318\n[08:38:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5865\n[08:38:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:38:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1627\nThe best R2 score of 0.5865\n--------------------------------------------------------------------------------\nV319\nFeature name V319\n[08:38:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.5710\n[08:38:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:38:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -2.5461\nThe best R2 score of -0.5710\n--------------------------------------------------------------------------------\nV320\nFeature name V320\n[08:39:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0435\n[08:39:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:39:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0688\n[08:39:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:39:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4031\n[08:39:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:39:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score -1.5642\nThe best R2 score of 0.4031\n--------------------------------------------------------------------------------\nV322\nFeature name V322\n[08:40:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7971\n[08:40:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:40:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9743\n[08:40:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:40:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7105\nThe best R2 score of 0.9743\n--------------------------------------------------------------------------------\nV323\nFeature name V323\n[08:40:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8190\n[08:40:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:40:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7932\nThe best R2 score of 0.8190\n--------------------------------------------------------------------------------\nV324\nFeature name V324\n[08:40:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8445\n[08:40:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:40:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9822\n[08:40:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:40:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9818\nThe best R2 score of 0.9822\n--------------------------------------------------------------------------------\nV329\nFeature name V329\n[08:40:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8342\n[08:40:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:40:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9247\n[08:40:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:40:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7653\nThe best R2 score of 0.9247\n--------------------------------------------------------------------------------\nV331\nFeature name V331\n[08:40:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -1.5959\n[08:40:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:40:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1643\n[08:41:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -16.7540\nThe best R2 score of 0.1643\n--------------------------------------------------------------------------------\nV332\nFeature name V332\n[08:41:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1210\n[08:41:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4019\n[08:41:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2026\nThe best R2 score of 0.4019\n--------------------------------------------------------------------------------\nV333\nFeature name V333\n[08:41:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0032\n[08:41:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5462\n[08:41:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2234\nThe best R2 score of 0.5462\n--------------------------------------------------------------------------------\nV337\nFeature name V337\n[08:41:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0623\n[08:41:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -3.1499\nThe best R2 score of 0.0623\n--------------------------------------------------------------------------------\nid_01\nFeature name id_01\n[08:41:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0657\n[08:41:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0973\n[08:41:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.0109\nThe best R2 score of 0.0973\n--------------------------------------------------------------------------------\nid_02\nFeature name id_02\n[08:41:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2925\n[08:41:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2662\nThe best R2 score of 0.2925\n--------------------------------------------------------------------------------\nid_03\nFeature name id_03\n[08:41:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0354\n[08:41:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1405\nThe best R2 score of -0.0354\n--------------------------------------------------------------------------------\nid_05\nFeature name id_05\n[08:41:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0579\n[08:41:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:41:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0397\nThe best R2 score of 0.0579\n--------------------------------------------------------------------------------\nid_06\nFeature name id_06\n[08:41:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0154\n[08:42:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:42:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0227\nThe best R2 score of 0.0154\n--------------------------------------------------------------------------------\nid_07\nFeature name id_07\n[08:42:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0483\n[08:42:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:42:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.2187\nThe best R2 score of 0.0483\n--------------------------------------------------------------------------------\nid_09\nFeature name id_09\n[08:42:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0154\n[08:42:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:42:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1018\nThe best R2 score of 0.0154\n--------------------------------------------------------------------------------\nid_31_chrome_version_newness\nFeature name id_31_chrome_version_newness\n[08:42:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0113\n[08:42:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:42:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0710\nThe best R2 score of 0.0113\n--------------------------------------------------------------------------------\nid_31_safari_version_newness\nFeature name id_31_safari_version_newness\n[08:42:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0087\n[08:42:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:42:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1185\nThe best R2 score of -0.0087\n--------------------------------------------------------------------------------\nid_33_resolution\nFeature name id_33_resolution\n[08:42:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0185\n[08:42:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:42:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0057\nThe best R2 score of 0.0185\n--------------------------------------------------------------------------------\nTransactionAmt_decimal\nFeature name TransactionAmt_decimal\n[08:42:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1644\n[08:42:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:42:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1886\n[08:42:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:42:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1770\nThe best R2 score of 0.1886\n--------------------------------------------------------------------------------\nTransactionAmt_decimal_length\nFeature name TransactionAmt_decimal_length\n[08:42:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3625\n[08:43:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:43:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3555\nThe best R2 score of 0.3625\n--------------------------------------------------------------------------------\nid_18_15.0\nFeature name id_18_15.0\n[08:43:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1568\n[08:43:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:43:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1608\n[08:43:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:43:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1589\nThe best R2 score of 0.1608\n--------------------------------------------------------------------------------\ncard3_143.0\nFeature name card3_143.0\n[08:44:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:44:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:44:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9697\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard3_144.0\nFeature name card3_144.0\n[08:44:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9816\n[08:44:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:44:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9681\nThe best R2 score of 0.9816\n--------------------------------------------------------------------------------\ncard3_150.0\nFeature name card3_150.0\n[08:44:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9946\n[08:44:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:44:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:44:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:45:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard3_185.0\nFeature name card3_185.0\n[08:45:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:45:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:45:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:45:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:45:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard3_infrequent_category\nFeature name card3_infrequent_category\n[08:45:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7029\n[08:45:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:45:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8709\n[08:46:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:46:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.7902\nThe best R2 score of 0.8709\n--------------------------------------------------------------------------------\nR_emaildomain_2_com\nFeature name R_emaildomain_2_com\n[08:46:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8689\n[08:46:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:46:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8770\n[08:46:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:46:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.8677\nThe best R2 score of 0.8770\n--------------------------------------------------------------------------------\nid_14_-420.0\nFeature name id_14_-420.0\n[08:46:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2087\n[08:47:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:47:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.7268\n[08:47:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:47:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.5869\nThe best R2 score of 0.7268\n--------------------------------------------------------------------------------\nid_14_-360.0\nFeature name id_14_-360.0\n[08:47:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2540\n[08:47:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:47:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6435\n[08:47:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:47:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.6226\nThe best R2 score of 0.6435\n--------------------------------------------------------------------------------\nid_14_60.0\nFeature name id_14_60.0\n[08:48:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0234\n[08:48:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:48:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0508\nThe best R2 score of -0.0234\n--------------------------------------------------------------------------------\nM4_M0\nFeature name M4_M0\n[08:48:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1491\n[08:48:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:48:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1541\n[08:48:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:48:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1393\nThe best R2 score of 0.1541\n--------------------------------------------------------------------------------\nM4_M1\nFeature name M4_M1\n[08:49:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0413\n[08:49:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:49:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0458\n[08:49:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0458\n--------------------------------------------------------------------------------\ncard6_credit\nFeature name card6_credit\n[08:49:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:49:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:49:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:49:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:49:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9998\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nM8_F\nFeature name M8_F\n[08:49:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2030\n[08:50:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:50:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2044\n[08:50:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:50:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1949\nThe best R2 score of 0.2044\n--------------------------------------------------------------------------------\nhours_1.0\nFeature name hours_1.0\n[08:50:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:50:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:50:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_2.0\nFeature name hours_2.0\n[08:50:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:51:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:51:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_3.0\nFeature name hours_3.0\n[08:51:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:51:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:51:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:51:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:51:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_4.0\nFeature name hours_4.0\n[08:51:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:51:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:51:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_5.0\nFeature name hours_5.0\n[08:52:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:52:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:52:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_6.0\nFeature name hours_6.0\n[08:52:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:52:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:52:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_11.0\nFeature name hours_11.0\n[08:52:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:52:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:52:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:53:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:53:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_12.0\nFeature name hours_12.0\n[08:53:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:53:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:53:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_13.0\nFeature name hours_13.0\n[08:53:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:53:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:53:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:53:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:53:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_14.0\nFeature name hours_14.0\n[08:54:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:54:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:54:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:54:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:54:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_15.0\nFeature name hours_15.0\n[08:54:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:54:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:54:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:54:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:54:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[08:54:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:55:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 1.0000\n[08:55:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:55:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n18, R2 score 1.0000\n[08:55:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:55:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n22, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_16.0\nFeature name hours_16.0\n[08:55:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:55:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:55:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_17.0\nFeature name hours_17.0\n[08:55:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:55:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:55:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:56:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:56:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[08:56:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:56:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_18.0\nFeature name hours_18.0\n[08:56:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:56:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:56:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_19.0\nFeature name hours_19.0\n[08:56:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:56:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:56:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_20.0\nFeature name hours_20.0\n[08:57:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:57:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:57:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_21.0\nFeature name hours_21.0\n[08:57:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:57:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:57:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nhours_22.0\nFeature name hours_22.0\n[08:57:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:57:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:57:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:58:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:58:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard4_american express\nFeature name card4_american express\n[08:58:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:58:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:58:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard4_discover\nFeature name card4_discover\n[08:58:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:58:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:58:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[08:58:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:58:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard4_mastercard\nFeature name card4_mastercard\n[08:59:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[08:59:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:59:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nid_13_19.0\nFeature name id_13_19.0\n[08:59:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1277\n[08:59:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:59:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5467\n[08:59:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[08:59:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2521\nThe best R2 score of 0.5467\n--------------------------------------------------------------------------------\nid_13_49.0\nFeature name id_13_49.0\n[09:00:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3470\n[09:00:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:00:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3577\n[09:00:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:00:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2748\nThe best R2 score of 0.3577\n--------------------------------------------------------------------------------\nid_13_52.0\nFeature name id_13_52.0\n[09:00:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3194\n[09:00:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:00:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3282\n[09:00:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:00:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.3141\nThe best R2 score of 0.3282\n--------------------------------------------------------------------------------\nR_emaildomain_anonymous.com\nFeature name R_emaildomain_anonymous.com\n[09:01:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1787\n[09:01:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:01:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2009\n[09:01:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:01:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1365\nThe best R2 score of 0.2009\n--------------------------------------------------------------------------------\nR_emaildomain_gmail.com\nFeature name R_emaildomain_gmail.com\n[09:01:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3332\n[09:01:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:01:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3532\n[09:02:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:02:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.3212\nThe best R2 score of 0.3532\n--------------------------------------------------------------------------------\nR_emaildomain_hotmail.com\nFeature name R_emaildomain_hotmail.com\n[09:02:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2827\n[09:02:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:02:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2766\nThe best R2 score of 0.2827\n--------------------------------------------------------------------------------\nP_emaildomain_4_com\nFeature name P_emaildomain_4_com\n[09:02:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1103\n[09:02:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:02:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1214\n[09:02:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:03:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1085\nThe best R2 score of 0.1214\n--------------------------------------------------------------------------------\nid_31_tablet_False\nFeature name id_31_tablet_False\n[09:03:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9076\n[09:03:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:03:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9236\n[09:03:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:03:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9158\nThe best R2 score of 0.9236\n--------------------------------------------------------------------------------\naddr2_60.0\nFeature name addr2_60.0\n[09:03:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9748\n[09:04:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:04:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9792\n[09:04:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:04:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9787\nThe best R2 score of 0.9792\n--------------------------------------------------------------------------------\naddr2_87.0\nFeature name addr2_87.0\n[09:04:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9994\n[09:04:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:04:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:04:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:04:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[09:04:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:04:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9997\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nid_19_100.0\nFeature name id_19_100.0\n[09:05:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0939\n[09:05:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:05:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1873\n[09:05:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:05:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0857\nThe best R2 score of 0.1873\n--------------------------------------------------------------------------------\nid_19_193.0\nFeature name id_19_193.0\n[09:05:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0315\n[09:05:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:05:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0510\n[09:05:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:05:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.0291\nThe best R2 score of 0.0510\n--------------------------------------------------------------------------------\nid_19_266.0\nFeature name id_19_266.0\n[09:06:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2601\n[09:06:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:06:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2420\nThe best R2 score of 0.2601\n--------------------------------------------------------------------------------\nid_19_271.0\nFeature name id_19_271.0\n[09:06:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0413\n[09:06:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:06:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0701\n[09:06:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:06:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score -0.0495\nThe best R2 score of 0.0701\n--------------------------------------------------------------------------------\nid_19_312.0\nFeature name id_19_312.0\n[09:06:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0700\n[09:07:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:07:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0721\n[09:07:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:07:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0251\nThe best R2 score of 0.0721\n--------------------------------------------------------------------------------\nid_19_321.0\nFeature name id_19_321.0\n[09:07:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0127\n[09:07:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:07:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0481\nThe best R2 score of 0.0127\n--------------------------------------------------------------------------------\nid_19_410.0\nFeature name id_19_410.0\n[09:07:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1033\n[09:07:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:07:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0919\nThe best R2 score of 0.1033\n--------------------------------------------------------------------------------\nid_19_427.0\nFeature name id_19_427.0\n[09:08:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1053\n[09:08:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:08:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0490\nThe best R2 score of 0.1053\n--------------------------------------------------------------------------------\nid_19_infrequent_category\nFeature name id_19_infrequent_category\n[09:08:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1600\n[09:08:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:08:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1584\nThe best R2 score of 0.1600\n--------------------------------------------------------------------------------\nM6_F\nFeature name M6_F\n[09:08:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2614\n[09:08:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:08:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2679\n[09:08:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:08:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2514\nThe best R2 score of 0.2679\n--------------------------------------------------------------------------------\nM3_F\nFeature name M3_F\n[09:09:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0559\n[09:09:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:09:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0494\nThe best R2 score of 0.0559\n--------------------------------------------------------------------------------\nM5_F\nFeature name M5_F\n[09:09:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1226\n[09:09:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:09:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1234\n[09:09:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:09:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1185\nThe best R2 score of 0.1234\n--------------------------------------------------------------------------------\ncard1_2884\nFeature name card1_2884\n[09:10:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:10:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:10:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9859\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard1_6019\nFeature name card1_6019\n[09:10:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9928\n[09:10:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:10:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9937\n[09:10:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:10:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9919\nThe best R2 score of 0.9937\n--------------------------------------------------------------------------------\ncard1_7508\nFeature name card1_7508\n[09:11:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9870\n[09:11:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:11:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:11:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:11:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard1_7585\nFeature name card1_7585\n[09:11:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9704\n[09:11:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:11:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9889\n[09:11:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:11:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9840\nThe best R2 score of 0.9889\n--------------------------------------------------------------------------------\ncard1_7919\nFeature name card1_7919\n[09:11:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9662\n[09:11:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:12:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9926\n[09:12:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:12:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[09:12:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:12:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9932\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard1_9500\nFeature name card1_9500\n[09:12:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9998\n[09:12:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:12:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9946\nThe best R2 score of 0.9998\n--------------------------------------------------------------------------------\ncard1_9633\nFeature name card1_9633\n[09:12:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9576\n[09:12:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:13:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9878\n[09:13:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:13:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9931\n[09:13:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:13:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9930\nThe best R2 score of 0.9931\n--------------------------------------------------------------------------------\ncard1_10616\nFeature name card1_10616\n[09:13:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9881\n[09:13:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:13:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9912\n[09:13:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:13:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9909\nThe best R2 score of 0.9912\n--------------------------------------------------------------------------------\ncard1_12544\nFeature name card1_12544\n[09:13:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9996\n[09:14:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:14:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:14:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:14:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9971\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard1_12695\nFeature name card1_12695\n[09:14:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:14:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:14:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9966\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard1_12839\nFeature name card1_12839\n[09:14:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9916\n[09:14:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:14:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9958\n[09:15:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:15:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[09:15:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:15:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard1_15066\nFeature name card1_15066\n[09:15:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9709\n[09:15:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:15:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:15:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:15:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard1_15885\nFeature name card1_15885\n[09:15:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9852\n[09:15:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:15:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9966\n[09:16:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:16:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9998\n[09:16:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:16:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9966\nThe best R2 score of 0.9998\n--------------------------------------------------------------------------------\ncard1_16132\nFeature name card1_16132\n[09:16:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9997\n[09:16:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:16:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9944\nThe best R2 score of 0.9997\n--------------------------------------------------------------------------------\ncard1_17188\nFeature name card1_17188\n[09:16:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9980\n[09:16:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:16:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9917\nThe best R2 score of 0.9980\n--------------------------------------------------------------------------------\ncard1_infrequent_category\nFeature name card1_infrequent_category\n[09:17:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6181\n[09:17:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:17:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9885\n[09:17:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:17:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9934\n[09:17:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:17:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9932\nThe best R2 score of 0.9934\n--------------------------------------------------------------------------------\nid_31_firefox_False\nFeature name id_31_firefox_False\n[09:18:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8548\n[09:18:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:18:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8661\n[09:18:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:18:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.8567\nThe best R2 score of 0.8661\n--------------------------------------------------------------------------------\nM2_F\nFeature name M2_F\n[09:18:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0264\n[09:18:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:18:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0212\nThe best R2 score of 0.0264\n--------------------------------------------------------------------------------\nP_emaildomain_anonymous.com\nFeature name P_emaildomain_anonymous.com\n[09:18:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0571\n[09:18:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:19:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0672\n[09:19:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:19:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.0538\nThe best R2 score of 0.0672\n--------------------------------------------------------------------------------\nP_emaildomain_aol.com\nFeature name P_emaildomain_aol.com\n[09:19:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0190\n[09:19:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:19:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0192\n[09:19:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0192\n--------------------------------------------------------------------------------\nP_emaildomain_bellsouth.net\nFeature name P_emaildomain_bellsouth.net\n[09:19:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0163\n[09:19:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:19:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0409\nThe best R2 score of 0.0163\n--------------------------------------------------------------------------------\nP_emaildomain_comcast.net\nFeature name P_emaildomain_comcast.net\n[09:19:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0120\n[09:20:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:20:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0088\nThe best R2 score of 0.0120\n--------------------------------------------------------------------------------\nP_emaildomain_cox.net\nFeature name P_emaildomain_cox.net\n[09:20:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0071\n[09:20:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:20:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0044\nThe best R2 score of 0.0071\n--------------------------------------------------------------------------------\nP_emaildomain_gmail.com\nFeature name P_emaildomain_gmail.com\n[09:20:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0282\n[09:20:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:20:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0293\n[09:20:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0293\n--------------------------------------------------------------------------------\nP_emaildomain_hotmail.com\nFeature name P_emaildomain_hotmail.com\n[09:20:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1507\n[09:20:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:20:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1537\n[09:21:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:21:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1120\nThe best R2 score of 0.1537\n--------------------------------------------------------------------------------\nP_emaildomain_icloud.com\nFeature name P_emaildomain_icloud.com\n[09:21:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0041\n[09:21:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:21:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0115\nThe best R2 score of -0.0041\n--------------------------------------------------------------------------------\nP_emaildomain_live.com\nFeature name P_emaildomain_live.com\n[09:21:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0130\n[09:21:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:21:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0600\nThe best R2 score of -0.0130\n--------------------------------------------------------------------------------\nP_emaildomain_live.com.mx\nFeature name P_emaildomain_live.com.mx\n[09:21:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0000\n[09:22:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:22:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0670\nThe best R2 score of -0.0000\n--------------------------------------------------------------------------------\nP_emaildomain_me.com\nFeature name P_emaildomain_me.com\n[09:22:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0018\n[09:22:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:22:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0377\nThe best R2 score of -0.0018\n--------------------------------------------------------------------------------\nP_emaildomain_msn.com\nFeature name P_emaildomain_msn.com\n[09:22:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0011\n[09:22:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:22:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0202\nThe best R2 score of -0.0011\n--------------------------------------------------------------------------------\nP_emaildomain_optonline.net\nFeature name P_emaildomain_optonline.net\n[09:22:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0083\n[09:22:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:22:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0203\nThe best R2 score of 0.0083\n--------------------------------------------------------------------------------\nP_emaildomain_outlook.com\nFeature name P_emaildomain_outlook.com\n[09:23:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0040\n[09:23:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:23:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0331\nThe best R2 score of 0.0040\n--------------------------------------------------------------------------------\nP_emaildomain_verizon.net\nFeature name P_emaildomain_verizon.net\n[09:23:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0076\n[09:23:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:23:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0178\n[09:23:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0178\n--------------------------------------------------------------------------------\nP_emaildomain_yahoo.com\nFeature name P_emaildomain_yahoo.com\n[09:23:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0302\n[09:23:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:23:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0298\nThe best R2 score of 0.0302\n--------------------------------------------------------------------------------\nP_emaildomain_infrequent_category\nFeature name P_emaildomain_infrequent_category\n[09:23:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0460\n[09:23:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:24:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0274\nThe best R2 score of 0.0460\n--------------------------------------------------------------------------------\naddr1_126.0\nFeature name addr1_126.0\n[09:24:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:24:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:24:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_181.0\nFeature name addr1_181.0\n[09:24:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:24:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:24:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:24:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:24:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_184.0\nFeature name addr1_184.0\n[09:24:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:25:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:25:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9984\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_191.0\nFeature name addr1_191.0\n[09:25:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:25:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:25:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_204.0\nFeature name addr1_204.0\n[09:25:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:25:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:25:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:25:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:25:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[09:25:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:26:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_231.0\nFeature name addr1_231.0\n[09:26:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9999\n[09:26:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:26:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9951\nThe best R2 score of 0.9999\n--------------------------------------------------------------------------------\naddr1_264.0\nFeature name addr1_264.0\n[09:26:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:26:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:26:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:26:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:26:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_272.0\nFeature name addr1_272.0\n[09:26:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:27:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:27:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:27:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:27:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_299.0\nFeature name addr1_299.0\n[09:27:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:27:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:27:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9995\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_310.0\nFeature name addr1_310.0\n[09:27:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9975\n[09:27:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:27:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:27:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:28:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_315.0\nFeature name addr1_315.0\n[09:28:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:28:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:28:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_325.0\nFeature name addr1_325.0\n[09:28:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:28:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:28:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_330.0\nFeature name addr1_330.0\n[09:28:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:28:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:28:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:29:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:29:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_337.0\nFeature name addr1_337.0\n[09:29:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:29:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:29:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_387.0\nFeature name addr1_387.0\n[09:29:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:29:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:29:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:29:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:29:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_433.0\nFeature name addr1_433.0\n[09:29:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:30:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:30:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:30:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:30:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9975\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_441.0\nFeature name addr1_441.0\n[09:30:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:30:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:30:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_472.0\nFeature name addr1_472.0\n[09:30:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:30:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:30:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_485.0\nFeature name addr1_485.0\n[09:30:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:31:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:31:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:31:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:31:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\naddr1_infrequent_category\nFeature name addr1_infrequent_category\n[09:31:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.7383\n[09:31:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:31:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9953\n[09:31:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:31:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9981\n[09:31:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:31:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9994\n[09:32:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:32:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n18, R2 score 0.9985\nThe best R2 score of 0.9994\n--------------------------------------------------------------------------------\nProductCD_H\nFeature name ProductCD_H\n[09:32:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:32:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:32:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nProductCD_R\nFeature name ProductCD_R\n[09:32:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:32:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:32:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nM9_F\nFeature name M9_F\n[09:32:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0367\n[09:33:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:33:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0374\n[09:33:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\nThe best R2 score of 0.0374\n--------------------------------------------------------------------------------\nid_20_222.0\nFeature name id_20_222.0\n[09:33:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5102\n[09:33:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:33:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6643\n[09:33:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:33:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.6281\nThe best R2 score of 0.6643\n--------------------------------------------------------------------------------\nid_20_325.0\nFeature name id_20_325.0\n[09:33:46] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1236\n[09:33:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:33:53] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0817\nThe best R2 score of 0.1236\n--------------------------------------------------------------------------------\nid_20_333.0\nFeature name id_20_333.0\n[09:34:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2135\n[09:34:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:34:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4834\n[09:34:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:34:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4474\nThe best R2 score of 0.4834\n--------------------------------------------------------------------------------\nid_20_401.0\nFeature name id_20_401.0\n[09:34:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0381\n[09:34:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:34:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0108\nThe best R2 score of 0.0381\n--------------------------------------------------------------------------------\nid_20_500.0\nFeature name id_20_500.0\n[09:34:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2413\n[09:34:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:34:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3849\n[09:35:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:35:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.1684\nThe best R2 score of 0.3849\n--------------------------------------------------------------------------------\nid_20_507.0\nFeature name id_20_507.0\n[09:35:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2923\n[09:35:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:35:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2630\nThe best R2 score of 0.2923\n--------------------------------------------------------------------------------\nid_20_533.0\nFeature name id_20_533.0\n[09:35:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5013\n[09:35:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:35:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6157\n[09:35:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:35:57] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.5932\nThe best R2 score of 0.6157\n--------------------------------------------------------------------------------\nid_20_549.0\nFeature name id_20_549.0\n[09:36:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3083\n[09:36:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:36:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.5867\n[09:36:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:36:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4860\nThe best R2 score of 0.5867\n--------------------------------------------------------------------------------\nid_20_595.0\nFeature name id_20_595.0\n[09:36:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1229\n[09:36:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:36:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0641\nThe best R2 score of 0.1229\n--------------------------------------------------------------------------------\nid_20_597.0\nFeature name id_20_597.0\n[09:37:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1220\n[09:37:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:37:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3940\n[09:37:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:37:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2551\nThe best R2 score of 0.3940\n--------------------------------------------------------------------------------\nid_20_612.0\nFeature name id_20_612.0\n[09:37:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0136\n[09:37:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:37:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0493\nThe best R2 score of 0.0136\n--------------------------------------------------------------------------------\nid_20_infrequent_category\nFeature name id_20_infrequent_category\n[09:37:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2680\n[09:37:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:37:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.4493\n[09:38:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:38:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.4184\nThe best R2 score of 0.4493\n--------------------------------------------------------------------------------\nid_38_F\nFeature name id_38_F\n[09:38:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.8476\n[09:38:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:38:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8682\n[09:38:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:38:43] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.8500\nThe best R2 score of 0.8682\n--------------------------------------------------------------------------------\ncard2_111.0\nFeature name card2_111.0\n[09:39:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:39:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:39:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:39:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:39:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_170.0\nFeature name card2_170.0\n[09:39:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:39:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:39:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_174.0\nFeature name card2_174.0\n[09:39:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:39:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:39:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_215.0\nFeature name card2_215.0\n[09:40:04] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:40:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:40:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_225.0\nFeature name card2_225.0\n[09:40:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:40:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:40:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:40:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:40:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[09:40:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:40:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_268.0\nFeature name card2_268.0\n[09:40:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9999\n[09:41:00] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:41:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9954\nThe best R2 score of 0.9999\n--------------------------------------------------------------------------------\ncard2_321.0\nFeature name card2_321.0\n[09:41:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:41:20] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:41:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_360.0\nFeature name card2_360.0\n[09:41:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9998\n[09:41:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:41:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:41:47] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:41:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9905\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_361.0\nFeature name card2_361.0\n[09:42:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:42:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:42:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:42:15] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:42:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_481.0\nFeature name card2_481.0\n[09:42:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:42:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:42:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_490.0\nFeature name card2_490.0\n[09:42:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:42:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:42:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:42:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:43:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_512.0\nFeature name card2_512.0\n[09:43:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:43:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:43:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:43:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:43:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_514.0\nFeature name card2_514.0\n[09:43:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:43:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:43:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_553.0\nFeature name card2_553.0\n[09:43:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:43:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:43:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:44:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:44:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_555.0\nFeature name card2_555.0\n[09:44:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:44:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:44:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_567.0\nFeature name card2_567.0\n[09:44:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9795\n[09:44:38] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:44:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:44:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:44:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9966\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_583.0\nFeature name card2_583.0\n[09:44:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:45:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:45:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard2_infrequent_category\nFeature name card2_infrequent_category\n[09:45:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6597\n[09:45:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:45:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9944\n[09:45:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:45:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9964\n[09:45:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:45:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9957\nThe best R2 score of 0.9964\n--------------------------------------------------------------------------------\nM7_F\nFeature name M7_F\n[09:46:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.3285\n[09:46:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:46:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3366\n[09:46:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:46:26] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.3268\nThe best R2 score of 0.3366\n--------------------------------------------------------------------------------\nweekday_1.0\nFeature name weekday_1.0\n[09:46:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:46:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:46:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:46:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:46:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[09:47:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:47:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nweekday_2.0\nFeature name weekday_2.0\n[09:47:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:47:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:47:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nweekday_3.0\nFeature name weekday_3.0\n[09:47:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:47:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:47:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nweekday_4.0\nFeature name weekday_4.0\n[09:47:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:47:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:47:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:48:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:48:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[09:48:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:48:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nweekday_5.0\nFeature name weekday_5.0\n[09:48:23] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:48:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:48:30] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard5_102.0\nFeature name card5_102.0\n[09:48:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:48:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:48:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:48:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:48:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard5_117.0\nFeature name card5_117.0\n[09:49:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:49:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:49:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:49:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:49:18] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9991\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard5_137.0\nFeature name card5_137.0\n[09:49:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:49:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:49:35] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard5_166.0\nFeature name card5_166.0\n[09:49:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9991\n[09:49:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:49:52] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:49:56] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:49:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[09:50:03] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:50:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 0.9996\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard5_195.0\nFeature name card5_195.0\n[09:50:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:50:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:50:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:50:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:50:28] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 1.0000\n[09:50:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:50:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n14, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard5_198.0\nFeature name card5_198.0\n[09:50:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9636\n[09:50:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:50:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.8697\nThe best R2 score of 0.9636\n--------------------------------------------------------------------------------\ncard5_219.0\nFeature name card5_219.0\n[09:51:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9993\n[09:51:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:51:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9967\nThe best R2 score of 0.9993\n--------------------------------------------------------------------------------\ncard5_224.0\nFeature name card5_224.0\n[09:51:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9998\n[09:51:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:51:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\n[09:51:32] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:51:34] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9997\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard5_226.0\nFeature name card5_226.0\n[09:51:45] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:51:49] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:51:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard5_229.0\nFeature name card5_229.0\n[09:52:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:52:05] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:52:08] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\ncard5_236.0\nFeature name card5_236.0\n[09:52:17] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 1.0000\n[09:52:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:52:24] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 1.0000\nThe best R2 score of 1.0000\n--------------------------------------------------------------------------------\nP_emaildomain_1_hotmail\nFeature name P_emaildomain_1_hotmail\n[09:52:33] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1532\n[09:52:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:52:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1491\nThe best R2 score of 0.1532\n--------------------------------------------------------------------------------\nP_emaildomain_1_outlook\nFeature name P_emaildomain_1_outlook\n[09:52:51] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0036\n[09:52:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:52:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.0216\nThe best R2 score of 0.0036\n--------------------------------------------------------------------------------\nP_emaildomain_1_infrequent_category\nFeature name P_emaildomain_1_infrequent_category\n[09:53:06] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0419\n[09:53:10] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:53:12] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0159\nThe best R2 score of 0.0419\n--------------------------------------------------------------------------------\nid_17_100.0\nFeature name id_17_100.0\n[09:53:21] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.5949\n[09:53:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:53:27] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.3455\nThe best R2 score of 0.5949\n--------------------------------------------------------------------------------\nid_17_166.0\nFeature name id_17_166.0\n[09:53:36] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.9128\n[09:53:40] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:53:42] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.9243\n[09:53:48] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:53:50] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.9141\nThe best R2 score of 0.9243\n--------------------------------------------------------------------------------\nid_32_24.0\nFeature name id_32_24.0\n[09:54:07] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.6282\n[09:54:11] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:54:14] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.6462\n[09:54:19] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:54:22] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.6217\nThe best R2 score of 0.6462\n--------------------------------------------------------------------------------\nDeviceInfo_MacOS\nFeature name DeviceInfo_MacOS\n[09:54:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.1444\n[09:54:41] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:54:44] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.1226\nThe best R2 score of 0.1444\n--------------------------------------------------------------------------------\nDeviceInfo_SM-J700M Build/MMB29K\nFeature name DeviceInfo_SM-J700M Build/MMB29K\n[09:54:54] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score -0.0075\n[09:54:58] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:55:01] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score -0.1016\nThe best R2 score of -0.0075\n--------------------------------------------------------------------------------\nDeviceInfo_Trident/7.0\nFeature name DeviceInfo_Trident/7.0\n[09:55:09] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.0626\n[09:55:13] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:55:16] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.0404\nThe best R2 score of 0.0626\n--------------------------------------------------------------------------------\nDeviceInfo_Windows\nFeature name DeviceInfo_Windows\n[09:55:25] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2669\n[09:55:29] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:55:31] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2696\n[09:55:37] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:55:39] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n10, R2 score 0.2296\nThe best R2 score of 0.2696\n--------------------------------------------------------------------------------\nDeviceInfo_infrequent_category\nFeature name DeviceInfo_infrequent_category\n[09:55:55] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n2, R2 score 0.2541\n[09:55:59] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n[09:56:02] WARNING: /workspace/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.\n6, R2 score 0.2495\nThe best R2 score of 0.2541\n--------------------------------------------------------------------------------\nCPU times: user 2h 57s, sys: 41min 6s, total: 2h 42min 3s\nWall time: 2h 41min 31s\n" ], [ "master_df.to_csv('master_df_time_adjusted_top_300.csv', index=False)", "_____no_output_____" ] ] ]

[ "code", "markdown", "code" ]

[ [ "code" ], [ "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "#installation, may not be needed\n#if (!requireNamespace(\"BiocManager\", quietly = TRUE))\n# install.packages(\"BiocManager\")\n\n#BiocManager::install(\"DESeq2\")", "_____no_output_____" ], [ "suppressMessages(library(DESeq2))", "_____no_output_____" ], [ "#UPF1 knockdown vs scramble\ncount_file <- \"colombo_combined_counts.tsv\"\nrow_names <- \"tx_id\"\ncountdata <- read.csv(count_file, row.names=row_names, sep='\\t')[ ,c('scr1_Homo_sapiens_RNA.Seq', 'scr2_Homo_sapiens_RNA.Seq', 'scr3_Homo_sapiens_RNA.Seq', 'UPF1_KD1_Homo_sapiens_RNA.Seq', 'UPF1_KD2_Homo_sapiens_RNA.Seq', 'UPF1_KD3_Homo_sapiens_RNA.Seq')]\ncountdata <- round(countdata)\ncondition <- factor(c(rep(\"ctl\", 3), rep(\"exp\", 3)))\n# Create a coldata frame and instantiate the DESeqDataSet. See ?DESeqDataSetFromMatrix\ncoldata <- data.frame(row.names=colnames(countdata), condition)\ndds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=~condition)\n# Run the DESeq pipeline\ndds <- DESeq(dds)\n#filter for at least 30 counts between all datasets\ndds <- dds[ rowSums(counts(dds)) > 30, ]\n# Get differential expression results\nres <- results(dds)\n## Order by adjusted p-value\nres <- res[order(res$padj), ]\n## Merge with normalized count data\n#res <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by=\"row.names\", sort=FALSE)\n#names(res)[1] <- \"Transcript\"\n## Write results\nwrite.table(res, file='DESeq_results/UPF1_kd_vs_scr.txt', quote=F, sep='\\t', col.names = NA, row.names = TRUE)\n\n#UPF1 rescue vs knockdown\ncountdata <- read.csv(count_file, row.names=row_names, sep='\\t')[ ,c('UPF1_KD1_Homo_sapiens_RNA.Seq', 'UPF1_KD2_Homo_sapiens_RNA.Seq', 'UPF1_KD3_Homo_sapiens_RNA.Seq', 'UPF1_rescue1_Homo_sapiens_RNA.Seq', 'UPF1_rescue2_Homo_sapiens_RNA.Seq', 'UPF1_rescue3_Homo_sapiens_RNA.Seq')]\ncountdata <- round(countdata)\ncondition <- factor(c(rep(\"ctl\", 3), rep(\"exp\", 3)))\n# Create a coldata frame and instantiate the DESeqDataSet. See ?DESeqDataSetFromMatrix\ncoldata <- data.frame(row.names=colnames(countdata), condition)\ndds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=~condition)\n# Run the DESeq pipeline\ndds <- DESeq(dds)\n#filter for at least 30 counts between all datasets\ndds <- dds[ rowSums(counts(dds)) > 30, ]\n# Get differential expression results\nres <- results(dds)\n## Order by adjusted p-value\nres <- res[order(res$padj), ]\n## Merge with normalized count data\n#res <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by=\"row.names\", sort=FALSE)\n#names(res)[1] <- \"Transcript\"\n## Write results\nwrite.table(res, file='DESeq_results/UPF1_rescue_vs_kd.txt', quote=F, sep='\\t', col.names = NA, row.names = TRUE)\n\n#SMG6 knockdown vs scramble\ncountdata <- read.csv(count_file, row.names=row_names, sep='\\t')[ ,c('scr1_Homo_sapiens_RNA.Seq', 'scr2_Homo_sapiens_RNA.Seq', 'scr3_Homo_sapiens_RNA.Seq', 'SMG6_KD1_Homo_sapiens_RNA.Seq', 'SMG6_KD2_Homo_sapiens_RNA.Seq', 'SMG6_KD3_Homo_sapiens_RNA.Seq')]\ncountdata <- round(countdata)\ncondition <- factor(c(rep(\"ctl\", 3), rep(\"exp\", 3)))\n# Create a coldata frame and instantiate the DESeqDataSet. See ?DESeqDataSetFromMatrix\ncoldata <- data.frame(row.names=colnames(countdata), condition)\ndds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=~condition)\n# Run the DESeq pipeline\ndds <- DESeq(dds)\n#filter for at least 30 counts between all datasets\ndds <- dds[ rowSums(counts(dds)) > 30, ]\n# Get differential expression results\nres <- results(dds)\n## Order by adjusted p-value\nres <- res[order(res$padj), ]\n## Merge with normalized count data\n#res <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by=\"row.names\", sort=FALSE)\n#names(res)[1] <- \"Transcript\"\n## Write results\nwrite.table(res, file='DESeq_results/SMG6_kd_vs_scr.txt', quote=F, sep='\\t', col.names = NA, row.names = TRUE)\n\n#SMG6 rescue vs knockdown\ncountdata <- read.csv(count_file, row.names=row_names, sep='\\t')[ ,c('SMG6_KD1_Homo_sapiens_RNA.Seq', 'SMG6_KD2_Homo_sapiens_RNA.Seq', 'SMG6_KD3_Homo_sapiens_RNA.Seq', 'SMG6_rescue1_Homo_sapiens_RNA.Seq', 'SMG6_rescue2_Homo_sapiens_RNA.Seq', 'SMG6_rescue3_Homo_sapiens_RNA.Seq')]\ncountdata <- round(countdata)\ncondition <- factor(c(rep(\"ctl\", 3), rep(\"exp\", 3)))\n# Create a coldata frame and instantiate the DESeqDataSet. See ?DESeqDataSetFromMatrix\ncoldata <- data.frame(row.names=colnames(countdata), condition)\ndds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=~condition)\n# Run the DESeq pipeline\ndds <- DESeq(dds)\n#filter for at least 30 counts between all datasets\ndds <- dds[ rowSums(counts(dds)) > 30, ]\n# Get differential expression results\nres <- results(dds)\n## Order by adjusted p-value\nres <- res[order(res$padj), ]\n## Merge with normalized count data\n#res <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by=\"row.names\", sort=FALSE)\n#names(res)[1] <- \"Transcript\"\n## Write results\nwrite.table(res, file='DESeq_results/SMG6_rescue_vs_kd.txt', quote=F, sep='\\t', col.names = NA, row.names = TRUE)\n\n#SMG7 knockdown vs scramble\ncountdata <- read.csv(count_file, row.names=row_names, sep='\\t')[ ,c('scr4_Homo_sapiens_RNA.Seq', 'scr5_Homo_sapiens_RNA.Seq', 'scr6_Homo_sapiens_RNA.Seq', 'SMG7_KD1_Homo_sapiens_RNA.Seq', 'SMG7_KD2_Homo_sapiens_RNA.Seq', 'SMG7_KD3_Homo_sapiens_RNA.Seq')]\ncountdata <- round(countdata)\ncondition <- factor(c(rep(\"ctl\", 3), rep(\"exp\", 3)))\n# Create a coldata frame and instantiate the DESeqDataSet. See ?DESeqDataSetFromMatrix\ncoldata <- data.frame(row.names=colnames(countdata), condition)\ndds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=~condition)\n# Run the DESeq pipeline\ndds <- DESeq(dds)\n#filter for at least 30 counts between all datasets\ndds <- dds[ rowSums(counts(dds)) > 30, ]\n# Get differential expression results\nres <- results(dds)\n## Order by adjusted p-value\nres <- res[order(res$padj), ]\n## Merge with normalized count data\n#res <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by=\"row.names\", sort=FALSE)\n#names(res)[1] <- \"Transcript\"\n## Write results\nwrite.table(res, file='DESeq_results/SMG7_kd_vs_scr.txt', quote=F, sep='\\t', col.names = NA, row.names = TRUE)\n\n#SMG7 rescue vs knockdown\ncountdata <- read.csv(count_file, row.names=row_names, sep='\\t')[ ,c('SMG7_KD1_Homo_sapiens_RNA.Seq', 'SMG7_KD2_Homo_sapiens_RNA.Seq', 'SMG7_KD3_Homo_sapiens_RNA.Seq', 'SMG7_rescue1_Homo_sapiens_RNA.Seq', 'SMG7_rescue2_Homo_sapiens_RNA.Seq', 'SMG7_rescue3_Homo_sapiens_RNA.Seq')]\ncountdata <- round(countdata)\ncondition <- factor(c(rep(\"ctl\", 3), rep(\"exp\", 3)))\n# Create a coldata frame and instantiate the DESeqDataSet. See ?DESeqDataSetFromMatrix\ncoldata <- data.frame(row.names=colnames(countdata), condition)\ndds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=~condition)\n# Run the DESeq pipeline\ndds <- DESeq(dds)\n#filter for at least 30 counts between all datasets\ndds <- dds[ rowSums(counts(dds)) > 30, ]\n# Get differential expression results\nres <- results(dds)\n## Order by adjusted p-value\nres <- res[order(res$padj), ]\n## Merge with normalized count data\n#res <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by=\"row.names\", sort=FALSE)\n#names(res)[1] <- \"Transcript\"\n## Write results\nwrite.table(res, file='DESeq_results/SMG7_rescue_vs_kd.txt', quote=F, sep='\\t', col.names = NA, row.names = TRUE)\n\n#SMG6/7 double_knockdown vs scramble\ncountdata <- read.csv(count_file, row.names=row_names, sep='\\t')[ ,c('scr4_Homo_sapiens_RNA.Seq', 'scr5_Homo_sapiens_RNA.Seq', 'scr6_Homo_sapiens_RNA.Seq', 'dKD1_Homo_sapiens_RNA.Seq', 'dKD2_Homo_sapiens_RNA.Seq', 'dKD3_Homo_sapiens_RNA.Seq')]\ncountdata <- round(countdata)\ncondition <- factor(c(rep(\"ctl\", 3), rep(\"exp\", 3)))\n# Create a coldata frame and instantiate the DESeqDataSet. See ?DESeqDataSetFromMatrix\ncoldata <- data.frame(row.names=colnames(countdata), condition)\ndds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=~condition)\n# Run the DESeq pipeline\ndds <- DESeq(dds)\n#filter for at least 30 counts between all datasets\ndds <- dds[ rowSums(counts(dds)) > 30, ]\n# Get differential expression results\nres <- results(dds)\n## Order by adjusted p-value\nres <- res[order(res$padj), ]\n## Merge with normalized count data\n#res <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by=\"row.names\", sort=FALSE)\n#names(res)[1] <- \"Transcript\"\n## Write results\nwrite.table(res, file='DESeq_results/dSMG_kd_vs_scr.txt', quote=F, sep='\\t', col.names = NA, row.names = TRUE)\n\n#SMG6 rescue vs SMG6/7 double knockdown\ncountdata <- read.csv(count_file, row.names=row_names, sep='\\t')[ ,c('dKD1_Homo_sapiens_RNA.Seq', 'dKD2_Homo_sapiens_RNA.Seq', 'dKD3_Homo_sapiens_RNA.Seq', 'dKD_SMG6_rescue1_Homo_sapiens_RNA.Seq', 'dKD_SMG6_rescue2_Homo_sapiens_RNA.Seq', 'dKD_SMG6_rescue3_Homo_sapiens_RNA.Seq')]\ncountdata <- round(countdata)\ncondition <- factor(c(rep(\"ctl\", 3), rep(\"exp\", 3)))\n# Create a coldata frame and instantiate the DESeqDataSet. See ?DESeqDataSetFromMatrix\ncoldata <- data.frame(row.names=colnames(countdata), condition)\ndds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=~condition)\n# Run the DESeq pipeline\ndds <- DESeq(dds)\n#filter for at least 30 counts between all datasets\ndds <- dds[ rowSums(counts(dds)) > 30, ]\n# Get differential expression results\nres <- results(dds)\n## Order by adjusted p-value\nres <- res[order(res$padj), ]\n## Merge with normalized count data\n#res <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by=\"row.names\", sort=FALSE)\n#names(res)[1] <- \"Transcript\"\n## Write results\nwrite.table(res, file='DESeq_results/SMG6_rescue_vs_dSMG_kd.txt', quote=F, sep='\\t', col.names = NA, row.names = TRUE)\n\n#SMG7 rescue vs SMG6/7 double knockdown\ncountdata <- read.csv(count_file, row.names=row_names, sep='\\t')[ ,c('dKD1_Homo_sapiens_RNA.Seq', 'dKD2_Homo_sapiens_RNA.Seq', 'dKD3_Homo_sapiens_RNA.Seq', 'dKD_SMG7_rescue1_Homo_sapiens_RNA.Seq', 'dKD_SMG7_rescue2_Homo_sapiens_RNA.Seq', 'dKD_SMG7_rescue3_Homo_sapiens_RNA.Seq')]\ncountdata <- round(countdata)\ncondition <- factor(c(rep(\"ctl\", 3), rep(\"exp\", 3)))\n# Create a coldata frame and instantiate the DESeqDataSet. See ?DESeqDataSetFromMatrix\ncoldata <- data.frame(row.names=colnames(countdata), condition)\ndds <- DESeqDataSetFromMatrix(countData=countdata, colData=coldata, design=~condition)\n# Run the DESeq pipeline\ndds <- DESeq(dds)\n#filter for at least 30 counts between all datasets\ndds <- dds[ rowSums(counts(dds)) > 30, ]\n# Get differential expression results\nres <- results(dds)\n## Order by adjusted p-value\nres <- res[order(res$padj), ]\n## Merge with normalized count data\n#res <- merge(as.data.frame(res), as.data.frame(counts(dds, normalized=TRUE)), by=\"row.names\", sort=FALSE)\n#names(res)[1] <- \"Transcript\"\n## Write results\nwrite.table(res, file='DESeq_results/SMG7_rescue_vs_dSMG_kd.txt', quote=F, sep='\\t', col.names = NA, row.names = TRUE)\n", "converting counts to integer mode\n\nestimating size factors\n\nestimating dispersions\n\ngene-wise dispersion estimates\n\nmean-dispersion relationship\n\nfinal dispersion estimates\n\nfitting model and testing\n\nconverting counts to integer mode\n\nestimating size factors\n\nestimating dispersions\n\ngene-wise dispersion estimates\n\nmean-dispersion relationship\n\nfinal dispersion estimates\n\nfitting model and testing\n\nconverting counts to integer mode\n\nestimating size factors\n\nestimating dispersions\n\ngene-wise dispersion estimates\n\nmean-dispersion relationship\n\nfinal dispersion estimates\n\nfitting model and testing\n\nconverting counts to integer mode\n\nestimating size factors\n\nestimating dispersions\n\ngene-wise dispersion estimates\n\nmean-dispersion relationship\n\nfinal dispersion estimates\n\nfitting model and testing\n\nconverting counts to integer mode\n\nestimating size factors\n\nestimating dispersions\n\ngene-wise dispersion estimates\n\nmean-dispersion relationship\n\nfinal dispersion estimates\n\nfitting model and testing\n\nconverting counts to integer mode\n\nestimating size factors\n\nestimating dispersions\n\ngene-wise dispersion estimates\n\nmean-dispersion relationship\n\nfinal dispersion estimates\n\nfitting model and testing\n\nconverting counts to integer mode\n\nestimating size factors\n\nestimating dispersions\n\ngene-wise dispersion estimates\n\nmean-dispersion relationship\n\nfinal dispersion estimates\n\nfitting model and testing\n\nconverting counts to integer mode\n\nestimating size factors\n\nestimating dispersions\n\ngene-wise dispersion estimates\n\nmean-dispersion relationship\n\nfinal dispersion estimates\n\nfitting model and testing\n\nconverting counts to integer mode\n\nestimating size factors\n\nestimating dispersions\n\ngene-wise dispersion estimates\n\nmean-dispersion relationship\n\nfinal dispersion estimates\n\nfitting model and testing\n\n" ] ] ]

[ "code" ]

[ [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "## Load the iris data", "_____no_output_____" ] ], [ [ "import matplotlib.pyplot as plt\n%matplotlib inline\nfrom sklearn.datasets import load_iris\nfrom numpy.linalg import inv\nimport pandas as pd\nimport numpy as np\n\niris = load_iris()\niris['data'][:5,:]\n\ny = np.where(iris['target'] == 2, 1, 0)\nX = iris['data']\n\nconst = np.ones(shape=y.shape).reshape(-1,1)\n\nmat = np.concatenate( (const, X), axis=1)\nmat[:5,:]", "_____no_output_____" ] ], [ [ "## Recall the algorithm we created for gradient descent for linear regression\nUsing the following cost function:\n$$J(w)=\\frac{1}{2}\\sum(y^{(i)} - \\hat{y}^{(i)})^2$$", "_____no_output_____" ] ], [ [ "import numpy as np\n\ndef gradientDescent(x, y, theta, alpha, m, numIterations):\n thetaHistory = list()\n \n xTrans = x.transpose()\n costList = list()\n \n for i in range(0, numIterations):\n # data x feature weights = y_hat\n hypothesis = np.dot(x, theta)\n # how far we are off\n loss = hypothesis - y \n # mse\n cost = np.sum(loss ** 2) / (2 * m)\n costList.append(cost)\n\n # avg gradient per example\n gradient = np.dot(xTrans, loss) / m \n\n # update\n theta = theta - alpha * gradient\n thetaHistory.append(theta)\n \n return thetaHistory, costList", "_____no_output_____" ] ], [ [ "## For Logistic regression we replace with our likehihood function:\n\n$$\nJ(w)=\\sum{[-y^{(i)}log(\\theta(z^{(i)}))-(1-y^{(i)})log(1-\\theta(z^{(i)})]}\n$$\n\n## And add the sigmoid function to bound $y$ between 0 and 1", "_____no_output_____" ] ], [ [ "def gradientDescent(x, y, alpha, numIterations):\n def mle(y,yhat):\n '''\n This replaces the mean squared error\n '''\n return (-y.dot(np.log(yhat)) - ((1-y)).dot(np.log(1-yhat)))\n \n def sigmoid(z):\n '''\n Transforms values to follow the sigmoid function and bound between 0 and 1\n '''\n return 1./(1. + np.exp(-np.clip(z, -250, 250)))\n \n # number of examples in the training data\n m = x.shape[0]\n\n # initialize weights to small random numbers\n theta = np.random.normal(loc=0.0, scale=0.1, size=x.shape[1])\n \n # history of theta values\n thetaHistory = list()\n \n xTrans = x.transpose()\n \n # history of cost values\n costList = list()\n \n for i in range(0, numIterations):\n \n # predicted value based on feature matrix and current weights\n hypothesis = np.dot(x, theta)\n \n # sigmoid transformation so we have bounded values\n hypothesis = sigmoid(hypothesis)\n \n # how far we are off from the actual value\n loss = hypothesis - y \n \n # determine cost based on the log likehilood function\n cost = mle(y, hypothesis)\n costList.append(cost)\n\n # avg gradient per example\n gradient = np.dot(xTrans, loss) / m \n\n # update the weights\n theta = theta - alpha * gradient\n thetaHistory.append(theta)\n \n return thetaHistory, costList", "_____no_output_____" ] ], [ [ "## Let's try it out\n- Run the algorithm, which gives us the weight and cost history. \n- Plot the cost to see if it converges. \n- Make predictions with the last batch of weights. \n- Apply the sigmoid function to the above predictions. \n- Plot the actual vs. predicted values. \n- Plot the evolution of the weights for each iteration.", "_____no_output_____" ] ], [ [ "iters = 500000\n\nimport datetime\n\nstart_ts = datetime.datetime.now()\nbetaHistory, costList = gradientDescent(mat, y, alpha=0.01, numIterations=iters)\n \nend_ts = datetime.datetime.now()\n\nprint(f'Completed in {end_ts-start_ts}')\n\n# cost history\nplt.plot(costList)\nplt.title(f'Final cost: {costList[-1]:,.2f}', loc='left')\nplt.show()\n\n# predict history\ngs_betas = betaHistory[iters-1]\ngs_predictions = np.dot(mat, gs_betas)\n\n# we need to apply the sigmoid/activation function to bound the predictions between (0,1)\ngs_predictions = 1./(1+np.exp(-gs_predictions))\n\nplt.plot(y, gs_predictions, 'bo', alpha=0.2)\nplt.xlabel('Actual')\nplt.ylabel('Predicted')\nplt.title('Gradient Descent Regression Fit on Training Data')\nplt.show()\n\nfrom collections import defaultdict\nthetas = defaultdict(list)\n\nfor i in range(len(betaHistory)):\n for j in range(len(betaHistory[i])):\n thetas[j].append(betaHistory[i][j])\n \nthetasD = pd.DataFrame.from_dict(thetas)\nthetasD.plot(legend=False)\nplt.title('Beta Estimates')\nplt.ylabel('Coefficient')\nplt.xlabel('Iteration')\nplt.show()", "Completed in 0:00:17.566409\n" ] ], [ [ "- We can see the loss is decreasing over the iterations. \n- Predictions are bounded between 0 and 1 because of the sigmoid function. \n- Weights update after each iteration and will eventually stabilize.", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# 셀레니움을 이용한 네이버 블로그(검색창) 크롤러\n- 네이버 메인 검색 페이지에서 크롤링한다.", "_____no_output_____" ] ], [ [ "import platform\nprint(platform.architecture())", "('64bit', '')\n" ], [ "!python --version", "Python 3.7.9\r\n" ], [ "pwd", "_____no_output_____" ], [ "# 네이버에서 검색어 입력받아 검색 한 후 블로그 메뉴를 선택하고\n# 오른쪽에 있는 검색옵션 버튼을 눌러서\n# 정렬 방식과 기간을 입력하기\n\n#Step 0. 필요한 모듈과 라이브러리를 로딩하고 검색어를 입력 받습니다.\nimport sys\nimport os\nimport pandas as pd\nimport numpy as np\nimport math\n\nfrom bs4 import BeautifulSoup\nimport requests\nimport urllib.request as req\n\nfrom selenium import webdriver\nfrom selenium.webdriver.common.keys import Keys\nimport time\nimport tqdm\nfrom tqdm.notebook import tqdm", "_____no_output_____" ], [ "query_txt = '성심당여행대전'\nstart_date= \"20190101\"\nend_date= \"20210501\"", "_____no_output_____" ], [ "os.getenv('HOME')", "_____no_output_____" ], [ "webdriver.__version__", "_____no_output_____" ], [ "#Step 1. 크롬 웹브라우저 실행\npath = os.getenv('HOME')+ '/chromedriver'", "_____no_output_____" ], [ "driver = webdriver.Chrome(path)\n# 사이트 주소는 네이버\nc\ntime.sleep(1)", "_____no_output_____" ], [ "#Step 2. 네이버 검색창에 \"검색어\" 검색\nelement = driver.find_element_by_name(\"query\")\nelement.send_keys(query_txt)\nelement.submit()\ntime.sleep(2)", "_____no_output_____" ], [ "#Step 3. \"블로그\" 카테고리 선택\ndriver.find_element_by_link_text(\"블로그\").click( ) \ntime.sleep(2)", "_____no_output_____" ], [ "#Step 4. 오른쪽의 검색 옵션 버튼 클릭\ndriver.find_element_by_class_name(\"btn_option._search_option_open_btn\").click( )\ntime.sleep(2)", "_____no_output_____" ], [ "driver.find_element_by_class_name(\"txt.txt_option._calendar_select_trigger\").click() # 관련도순 xpath\n# element.find_element_by_css_selector(\"#header > div.header_common > div > div.area_search > form > fieldset > a.button.button_blog\").click() # 관련도순 xpath\n# element.clear()\n# element.send_keys(query_txt) # query_txt는 위에서 입력한 '이재용'\n# element.submit()", "_____no_output_____" ], [ "#Step 1. 크롬 웹브라우저 실행\npath = os.getenv('HOME')+ '/chromedriver'\n\ndriver = webdriver.Chrome(path)\n# 사이트 주소는 네이버\ndriver.get('http://www.naver.com')\ntime.sleep(0.1)\n\n# # login\n# login = {\n# \"id\" : \"iminu95\",\n# \"pw\" : \"95bbkmjamy\"\n# }\n\n# # 아이디와 비밀번호를 입력합니다.\n# time.sleep(0.5) ## 0.5초\n\n# driver.find_element_by_class_name('link_login').click( )\n# time.sleep(1)\n\n# # driver.find_element_by_name('id').send_keys('아이디') # \"아이디라는 값을 보내준다\"\n# driver.find_element_by_name('id').send_keys(login.get(\"id\"))\n# time.sleep(0.5) ## 0.5초\n# driver.find_element_by_name('pw').send_keys(login.get(\"pw\")) \n# time.sleep(0.5) ## 0.5초\n# driver.find_element_by_class_name('btn_global').click( )\n# time.sleep(0.5) ## 0.5초\n\n\n\n\n#Step 2. 네이버 검색창에 \"검색어\" 검색\nelement = driver.find_element_by_name(\"query\")\nelement.send_keys(query_txt)\nelement.submit()\ntime.sleep(0.1)\n\n#Step 3. \"블로그\" 카테고리 선택\ndriver.find_element_by_link_text(\"블로그\").click( ) \ntime.sleep(2)\n\n#Step 4. 오른쪽의 검색 옵션 버튼 클릭\ndriver.find_element_by_class_name(\"btn_option._search_option_open_btn\").click( )\ntime.sleep(2)\n\n\n#Step 6. 날짜 입력\n# driver.find_element_by_class_name(\"txt.txt_option._calendar_select_trigger\").click() # 관련도순 xpath\n\n# driver.find_element_by_id(\"search_start_date\").send_keys(start_date)\n# driver.find_element_by_id(\"search_end_date\").send_keys(end_date)\n# time.sleep(0.1)\n\n# driver.find_element_by_id(\"periodSearch\").click()\n# time.sleep(0.1)\n\n# searched_post_num = driver.find_element_by_class_name('search_number').text\n# print(searched_post_num)\n\nurl_list = []\ntitle_list = []\n\ntotal_page = 2 \n# total_page = math.ceil(int(searched_post_num.replace(',', '').strip('건')) / 7)\nprint('total_page :', total_page)\n\nfor i in tqdm(range(0, total_page)): # 페이지 번호\n url = f'https://section.blog.naver.com/Search/Post.naver?pageNo={i}&rangeType=sim&orderBy=recentdate&startDate={start_date}&endDate={end_date}&keyword={query_txt}'\n driver.get(url)\n# response = requests.get(url)\n# soup = BeautifulSoup(response.text, 'html.parser')\n# print(soup)\n time.sleep(0.5)\n# area = soup.findAll('div', {'class' : 'list_search_post'}) #.find_all('a', {'class' : 'url'})\n# print(area)\n \n # URL 크롤링 시작\n titles = \"a.sh_blog_title._sp_each_url._sp_each_title\" # #content\n article_raw = driver.find_elements_by_class_name(titles)\n# article_raw = driver.find_elements_by_css_selector('#content > section > div.area_list_search > div:nth-child(1)')\n \n# article_raw = driver.find_elements_by_xpath(f'//*[@id=\"content\"]/section/div[2]/div[{i}]')\n \n# print(article_raw)\n\n # url 크롤링 시작 # 7개 \n for article in article_raw:\n url = article.get_attribute('href') \n print(url)\n url_list.append(url)\n \n # 제목 크롤링 시작 \n for article in article_raw:\n title = article.get_attribute('title') \n title_list.append(title)\n \n print(title)\n \nprint('url갯수: ', len(url_list))\nprint('url갯수: ', len(title_list))\n\n# df = pd.DataFrame({'url':url_list, 'title':title_list})\n\n# # 저장하기\n# df.to_csv(\"./blog_url.csv\")", "total_page : 2\n" ], [ "li = [2, 3, 4, 4, 5, 6, 7, 8]\nlen(li)", "_____no_output_____" ], [ "for i in range(0, 8, 2):\n print(i)", "0\n2\n4\n6\n" ], [ "new = []\nfor i in range(0, len(li)-1, 2):\n new.append([li[i], li[i+1]])", "_____no_output_____" ], [ "new", "_____no_output_____" ], [ "article_raw = driver.find_elements_by_xpath('//*[@id=\"content\"]/section/div[2]/div[1]')\n# article_raw.get_attribute('href')\nfor i in article_raw:\n print(i.get_attribute('href'))", "None\n" ], [ "//*[@id=\"content\"]/section/div[2]\n//*[@id=\"content\"]/section/div[2]\n//*[@id=\"content\"]/section/div[2]\n\n//*[@id=\"content\"]/section/div[2]/div[1]\n//*[@id=\"content\"]/section/div[2]/div[2]\n//*[@id=\"content\"]/section/div[2]/div[3]\n...\n//*[@id=\"content\"]/section/div[2]/div[7]", "_____no_output_____" ] ], [ [ "1 page = 7 posts\n72 page search\n\nsample = https://section.blog.naver.com/Search/Post.naver?pageNo=1&rangeType=PERIOD&orderBy=sim&startDate=2019-01-01&endDate=2021-05-01&keyword=%EC%84%B1%EC%8B%AC%EB%8B%B9%EC%97%AC%ED%96%89%EB%8C%80%EC%A0%84", "_____no_output_____" ] ], [ [ "## 제목 눌러서 블로그 페이지 열기\ndriver.find_element_by_class_name('title').click()\ntime.sleep(1)", "_____no_output_____" ], [ "type(searched_post_num), searched_post_num", "_____no_output_____" ], [ "import re", "_____no_output_____" ], [ "re.sub('^[0-9]', '', searched_post_num)", "_____no_output_____" ], [ "searched_post_num", "_____no_output_____" ], [ "searched_post_num.replace(',', '').replace('건', '')", "_____no_output_____" ], [ "total_page = math.ceil(int(searched_post_num.replace(',', '').strip('건')) / 7)\ntotal_page", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "from google.colab import drive\ndrive.mount('/content/drive')", "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly&response_type=code\n\nEnter your authorization code:\n··········\nMounted at /content/drive\n" ], [ "%run \"./Unet/train.py\"", "The device being used is: cuda\n\nTraining Started.....\n" ], [ "!cp -r \"./Unet\" \"./drive/My Drive/\"", "_____no_output_____" ], [ "%run \"./Unet/test.py\"", "_____no_output_____" ], [ "", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import numpy as np", "_____no_output_____" ], [ "list = [0,1,2,3,4,5,6,7,8]", "_____no_output_____" ], [ "len(list)", "_____no_output_____" ], [ "if len(list) < 9:\n try:\n num = int(\"error\")\n except ValueError:\n raise ValueError(\"List must contain nine numbers.\") ", "_____no_output_____" ], [ "calculations={}\na = np.array(list)\na = a.reshape(3,3)\na", "_____no_output_____" ] ], [ [ "{\n 'mean': [axis1, axis2, flattened],\n \n 'variance': [axis1, axis2, flattened],\n \n 'standard deviation': [axis1, axis2, flattened],\n \n 'max': [axis1, axis2, flattened],\n \n 'min': [axis1, axis2, flattened],\n \n 'sum': [axis1, axis2, flattened]\n}", "_____no_output_____" ] ], [ [ "calculations['mean']= [a.mean(axis=0).tolist(), a.mean(axis=1).tolist(), a.mean().tolist()]\ncalculations['mean']", "_____no_output_____" ], [ "calculations['variance']= [a.var(axis=0).tolist(), a.var(axis=1).tolist(), a.var().tolist()]\ncalculations", "_____no_output_____" ], [ "calculations['standard deviation']= [a.std(axis=0).tolist(), a.std(axis=1).tolist(), a.std().tolist()]\ncalculations", "_____no_output_____" ], [ "calculations['max']= [a.max(axis=0).tolist(), a.max(axis=1).tolist(), a.max().tolist()]\ncalculations['min']= [a.min(axis=0).tolist(), a.min(axis=1).tolist(), a.min().tolist()]\ncalculations['sum']= [a.sum(axis=0).tolist(), a.sum(axis=1).tolist(), a.sum().tolist()]\ncalculations", "_____no_output_____" ] ] ]

[ "code", "markdown", "code" ]

[ [ "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/abhibhaw/Torrent-To-Google-Drive-Downloader/blob/master/Torrent_To_Google_Drive_Downloader.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ], [ "# Torrent To Google Drive Downloader ", "_____no_output_____" ], [ "**Important Note:** To get more disk space:\n> Go to Runtime -> Change Runtime and give GPU as the Hardware Accelerator. You will get around 384GB to download any torrent you want.", "_____no_output_____" ], [ "### Install libtorrent and Initialize Session", "_____no_output_____" ] ], [ [ "!apt install python3-libtorrent\n\nimport libtorrent as lt\n\nses = lt.session()\nses.listen_on(6881, 6891)\ndownloads = []", "_____no_output_____" ] ], [ [ "### Mount Google Drive\nTo stream files we need to mount Google Drive.", "_____no_output_____" ] ], [ [ "from google.colab import drive\n\ndrive.mount(\"/content/drive\")", "_____no_output_____" ] ], [ [ "### Add From Torrent File\nYou can run this cell to add more files as many times as you want", "_____no_output_____" ] ], [ [ "from google.colab import files\n\nsource = files.upload()\nparams = {\n \"save_path\": \"/content/drive/My Drive/Torrent\",\n \"ti\": lt.torrent_info(list(source.keys())[0]),\n}\ndownloads.append(ses.add_torrent(params))", "_____no_output_____" ] ], [ [ "### Add From Magnet Link\nYou can run this cell to add more files as many times as you want", "_____no_output_____" ] ], [ [ "params = {\"save_path\": \"/content/drive/My Drive/Torrent\"}\n\nwhile True:\n magnet_link = input(\"Enter Magnet Link Or Type Exit: \")\n if magnet_link.lower() == \"exit\":\n break\n downloads.append(\n lt.add_magnet_uri(ses, magnet_link, params)\n )\n", "_____no_output_____" ] ], [ [ "### Start Download\nSource: https://stackoverflow.com/a/5494823/7957705 and [#3 issue](https://github.com/FKLC/Torrent-To-Google-Drive-Downloader/issues/3) which refers to this [stackoverflow question](https://stackoverflow.com/a/6053350/7957705)", "_____no_output_____" ] ], [ [ "import time\nfrom IPython.display import display\nimport ipywidgets as widgets\n\nstate_str = [\n \"queued\",\n \"checking\",\n \"downloading metadata\",\n \"downloading\",\n \"finished\",\n \"seeding\",\n \"allocating\",\n \"checking fastresume\",\n]\n\nlayout = widgets.Layout(width=\"auto\")\nstyle = {\"description_width\": \"initial\"}\ndownload_bars = [\n widgets.FloatSlider(\n step=0.01, disabled=True, layout=layout, style=style\n )\n for _ in downloads\n]\ndisplay(*download_bars)\n\nwhile downloads:\n next_shift = 0\n for index, download in enumerate(downloads[:]):\n bar = download_bars[index + next_shift]\n if not download.is_seed():\n s = download.status()\n\n bar.description = \" \".join(\n [\n download.name(),\n str(s.download_rate / 1000),\n \"kB/s\",\n state_str[s.state],\n ]\n )\n bar.value = s.progress * 100\n else:\n next_shift -= 1\n ses.remove_torrent(download)\n downloads.remove(download)\n bar.close() # Seems to be not working in Colab (see https://github.com/googlecolab/colabtools/issues/726#issue-486731758)\n download_bars.remove(bar)\n print(download.name(), \"complete\")\n time.sleep(1)\n", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Analysis of enrichment", "_____no_output_____" ] ], [ [ "import glob\n\nimport json\nimport math\n\nimport numpy as np\nimport pandas as pd\nimport seaborn as sns\nimport matplotlib.pyplot as plt\n\nfrom functools import reduce\nfrom collections import OrderedDict, defaultdict\nfrom sklearn.feature_extraction.text import TfidfVectorizer\nfrom scipy.stats import fisher_exact as fisher\nfrom scipy.stats import chi2_contingency as chisq", "_____no_output_____" ], [ "def ease(n_outliers_path, n_total_path, n_outliers, n_total):\n \"\"\"\n Calculates a contingency table EASE score\n [x y]\n [z k]\n :param n_in_path: number of outliers in the pathway\n :param n_total_path: total number of genes in the pathway\n :param n_outliers: total number of outliers\n :param n_total: total number of genes analysed\n :return:\n \"\"\"\n\n x = max(0, n_outliers_path - 1) # in category, enriched\n y = n_total_path # total, enriched\n z = n_outliers - n_outliers_path # in category, not enriched\n k = n_total - n_total_path # total, not enriched\n \n #if x <= 10:\n _, pvalue = fisher(([[x, y], [z, k]]), alternative='greater')\n #else:\n # _, pvalue, _, _ = chisq(([[x, y], [z, k]]))\n\n return pvalue", "_____no_output_____" ] ], [ [ "## Collecting all pathway names", "_____no_output_____" ] ], [ [ "pathway_tables = glob.glob(\"../pathways/*/gp.csv\")\ndfs = [pd.read_csv(table) for table in pathway_tables]\nfor i, df in enumerate(dfs):\n dfs[i] = df.set_index(\"SYMBOL\")\n dfs[i].sort_index(inplace=True)\n #print(dfs[i].shape)\ndfs[0]\nall_entries = list(pd.concat(dfs, axis=1, sort=True).columns)", "_____no_output_____" ], [ "all_entries[0:10]", "_____no_output_____" ], [ "structures = pd.read_csv(\"../extracted/classification_pathways.csv\", header=0, index_col=\"Pathway\")\nstructures = pd.DataFrame(structures, dtype=bool)", "_____no_output_____" ], [ "del structures[\"DUPLICATE?\"], structures[\"TRUTHFULNESS\"], structures[\"Garbage\"]\nstructures.head()", "_____no_output_____" ], [ "all_2 = set(structures.index)\nset(all_entries) - all_2", "_____no_output_____" ], [ "pathway_types = dict()\nfor pathway in sorted(all_entries):\n x = structures.loc[pathway]\n pathway_types[pathway] = x[x].index[0]", "_____no_output_____" ], [ "reverse_counter = defaultdict(int)\nfor pathway in sorted(all_entries):\n category = pathway_types[pathway]\n reverse_counter[category] += 1\nreverse_counter", "_____no_output_____" ], [ "ALL_PATHS = sum(reverse_counter.values())\nALL_PATHS", "_____no_output_____" ] ], [ [ "# By histone tag:", "_____no_output_____" ] ], [ [ "my_tags = [\"H3K4me3\", \"H3K9ac\", \"H3K27ac\", \"H3K27me3\", \"H3K9me3\"]", "_____no_output_____" ], [ "ENR_COUNTERS = dict()\nfor hg_tag in my_tags:\n files_up_human = glob.glob(f\"../extracted/Human_{hg_tag}_pathways_up*\")\n files_down_human = glob.glob(f\"../extracted/Human_{hg_tag}_pathways_down*\")\n files_up_mouse = glob.glob(f\"../extracted/Mouse_{hg_tag}_pathways_up*\")\n files_down_mouse = glob.glob(f\"../extracted/Mouse_{hg_tag}_pathways_down*\")\n\n files = {\"Human+\": files_up_human[0],\n \"Human-\": files_down_human[0],\n \"Mouse+\": files_up_mouse[0],\n \"Mouse-\": files_down_mouse[0]}\n\n enriched_counter = defaultdict(lambda: defaultdict(int))\n for xtype in files:\n with open(files[xtype], \"r\") as file:\n en_pathways = file.read().strip().split(\"\\n\")\n for pw in en_pathways:\n cat = pathway_types[pw]\n enriched_counter[xtype][cat] += 1\n enriched_counter = pd.DataFrame(enriched_counter).T.fillna(0)\n enriched_counter = pd.DataFrame(enriched_counter, dtype=int)\n ENR_COUNTERS[hg_tag] = enriched_counter\nENR_COUNTERS[my_tags[0]]", "_____no_output_____" ] ], [ [ "Calculates a contingency table EASE score \n[x y] \n[z k] \n:param n_in_path: number of outliers in the pathway \n:param n_total_path: total number of genes in the pathway \n:param n_outliers: total number of outliers \n:param n_total: total number of genes analysed \n:return: ", "_____no_output_____" ] ], [ [ "ksi = defaultdict(dict)\nsigns = {\"+\": \"positively\\u00A0enriched\\u00A0(+)\",\n \"-\": \"negatively\\u00A0enriched\\u00A0(-)\"}\nfor hg_tag in my_tags:\n enriched_counter = ENR_COUNTERS[hg_tag]\n for sign in [\"+\", \"-\"]:\n for org in [\"Human\", \"Mouse\"]:\n for category in enriched_counter:\n n1 = enriched_counter[category][f\"{org}{sign}\"]\n n2 = sum(enriched_counter.loc[f\"{org}{sign}\"])\n n3 = reverse_counter[category]\n n4 = ALL_PATHS\n #print(n1, n2, n3, n4)\n ksi[category][f\"{org},\\u00A0{hg_tag},\\u00A0{signs[sign]}\"] = ease(n1, n2, n3, n4)\npd.DataFrame(ksi).to_csv(f\"../extracted/pvalues.csv\")\npd.DataFrame(ksi)", "_____no_output_____" ], [ "TAU = pd.DataFrame(ksi)", "_____no_output_____" ], [ "def get_highlighter_min(color, point):\n def highlight_min(s):\n '''\n highlight the minimums in a Series.\n '''\n is_max = s <= point\n return [f'background-color: {color}' if v else '' for v in is_max]\n return highlight_min", "_____no_output_____" ], [ "data_round = np.round(TAU, 3)\ncm = sns.light_palette(\"green\", as_cmap=True, reverse=True)\n\ns = data_round.style.apply(get_highlighter_min(\"green\", 0.05), subset=([i for i in TAU.index if \"+\" in i], TAU.columns))\ncm = sns.light_palette(\"red\", as_cmap=True, reverse=True)\n\ns.apply(get_highlighter_min(\"red\", 0.05), subset=([i for i in TAU.index if \"-\" in i], TAU.columns))\ns", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# AWS Rekognition to get bbox\nimport numpy as np\nimport boto3\nfrom PIL import Image, ImageDraw, ExifTags, ImageColor, ImageFont\nfrom matplotlib import pyplot as plt\nfrom utils.rekognition import determine_color, draw_animal_count\nimport cv2\nimport time\nimport math\nimport os\nfrom utils.config import *\nfrom utils.fix_annotation import *", "_____no_output_____" ], [ "def save_frames(src_file, output_path, output_video, fps=cv2.CAP_PROP_FPS):\n start = time.time()\n \n cap = cv2.VideoCapture(src_file)\n frameRate = cap.get(fps) #frame rate\n print('frameRate', frameRate)\n #function to write a video\n height = cap.get(cv2.CAP_PROP_FRAME_HEIGHT)\n width = cap.get(cv2.CAP_PROP_FRAME_WIDTH)\n \n imgSize = (int(width), int(height))\n \n fourcc = cv2.VideoWriter_fourcc(*'mp4v')\n videoWriter = cv2.VideoWriter(output_video, fourcc, frameRate, imgSize)\n \n while(cap.isOpened()):\n frameId = cap.get(1) #current frame number\n \n ret, frame = cap.read()\n \n if (ret != True):\n break\n else:\n #inference on the extracted frame\n hasFrame, imageBytes = cv2.imencode(\".jpg\", frame)\n\n if(hasFrame):\n # creating image object of array\n data = Image.fromarray(frame)\n # shrink the frame image size\n # saving a JPG file\n data.save(output_path+'{}.jpg'.format(int(frameId)))\n \n if frameId % 50 == 0:\n print(\"Finish Processing {} frames\".format(int(frameId)))\n lap = time.time()\n print('lap time: ', lap - start)\n videoWriter.write(frame)\n\n cap.release()\n videoWriter.release()\n cv2.destroyAllWindows()\n \n end = time.time()\n print('total time lapse', end - start)\n#'cattle_single_1', 'cattle_multi_1'\nvideo_name_list = ['tests']\nvideo_format = ['.mp4']\nfor v_idx, video_name in enumerate(video_name_list):\n src_video = 'video_data/input_video/'+video_name+video_format[v_idx]\n output_img_path = 'frame_img/'+video_name+'/'\n if not os.path.exists(output_img_path):\n os.makedirs(output_img_path)\n output_video = output_img_path+'ori_video.mp4'\n print(output_video)\n save_frames(src_video, output_img_path, output_video)\n print()", "frame_img/tests/ori_video.mp4\nframeRate 30.0\nFinish Processing 0 frames\nlap time: 0.4605433940887451\nFinish Processing 50 frames\nlap time: 8.51185417175293\nFinish Processing 100 frames\nlap time: 16.091991424560547\nFinish Processing 150 frames\nlap time: 23.63861584663391\nFinish Processing 200 frames\nlap time: 31.303373098373413\nFinish Processing 250 frames\nlap time: 38.85345196723938\nFinish Processing 300 frames\nlap time: 46.18950009346008\nFinish Processing 350 frames\nlap time: 53.64163899421692\ntotal time lapse 59.122246980667114\n\n" ] ] ]

[ "code" ]

[ [ "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Basic usage", "_____no_output_____" ], [ "Thunder offers a variety of analyses and workflows for spatial and temporal data. When run on a cluster, most methods are efficiently and automatically parallelized, but Thunder can also be used on a single machine, especially for testing purposes. \n\nWe'll walk through a very simple example here as an introduction. The entry point for most workflows is the ``ThunderContext``. If you type ``thunder`` to start the interactive shell, this context is automatically provided as ``tsc``, which is an object that primarily provides functionality for loading and exporting data.\n\nWe'll start by loading and exploring some toy example data:", "_____no_output_____" ] ], [ [ "data = tsc.loadExample('fish-series')", "_____no_output_____" ] ], [ [ "``data`` is a ``Series`` object, which is a generic collection of one-dimensional array data sharing a common index. We can inspect it to see metadata:", "_____no_output_____" ] ], [ [ "data", "_____no_output_____" ] ], [ [ "A ``Series`` object is a collection of key-value records, each containing an identifier as a key and a one-dimensional array as a value. We can look at the first key and value by using ``first()``.", "_____no_output_____" ] ], [ [ "key, value = data.first()", "_____no_output_____" ] ], [ [ "We see that the ``first`` key in this example ``Series`` data is the tuple (0,0,0), corresponding to an x, y, z coordinate of an original movie.", "_____no_output_____" ] ], [ [ "key", "_____no_output_____" ] ], [ [ "The value in this case is a time series of 240 observations, represented as a 1d numpy array.\n", "_____no_output_____" ] ], [ [ "value.shape", "_____no_output_____" ] ], [ [ "We can extract a random subset of records and plot their time series, after converting to `TimeSeries` (which enables time-specific methods), and applying a simple baseline normalization. Here and elsewhere, we'll use the excellent ``seaborn`` package for styling figures, but this is entirely optional.", "_____no_output_____" ] ], [ [ "%matplotlib inline", "_____no_output_____" ], [ "import matplotlib.pyplot as plt\nimport seaborn as sns\nsns.set_context(\"notebook\")", "_____no_output_____" ], [ "examples = data.toTimeSeries().normalize().subset(50, thresh=0.05)\nsns.set_style('darkgrid')\nplt.plot(examples.T);", "_____no_output_____" ] ], [ [ "We can also compute a statistic for each record using the method:", "_____no_output_____" ] ], [ [ "means = data.seriesStdev()\nmeans.first()", "_____no_output_____" ] ], [ [ "``means`` is now itself a ``Series``, where the value of each record is the mean across time\n", "_____no_output_____" ], [ "For this ``Series``, since the keys correspond to spatial coordinates, we can ``pack`` the results back into a local array. ``pack`` is an operation that converts ``Series`` data, with spatial coordinates as keys, into an n-dimensional numpy array. In this case, the result is 3D, reflecting the original input data.", "_____no_output_____" ] ], [ [ "img = means.pack()\nimg.shape", "_____no_output_____" ] ], [ [ "``pack`` is an example of a local operation, meaning that all the data involved will be sent to the Spark driver node. For larger data sets, this can be very problematic - it's a good idea to downsample, subselect, or otherwise reduce the size of your data before attempting to ``pack`` large data sets!\n\nTo look at this array as an image, we'll use `matplotlib` via a helper function included with Thunder.", "_____no_output_____" ] ], [ [ "from thunder import Colorize\nimage = Colorize.image\nimage(img[:,:,0])", "_____no_output_____" ] ], [ [ "It's also easy to export the result to a ``numpy`` or ``MAT`` file.", "_____no_output_____" ], [ "```\ntsc.export(img, \"directory\", \"npy\")\ntsc.export(img, \"directory\", \"mat\")\n```", "_____no_output_____" ], [ "This will put a ``npy`` file or ``MAT`` file called ``meanval`` in the folder ``directory`` in your current directory. You can also export to a location of Amazon S3 or Google Storage if path is specified with an `s3n://`or `gs://` prefix.", "_____no_output_____" ], [ "Thunder includes several other toy data sets, to see the available ones:", "_____no_output_____" ] ], [ [ "tsc.loadExample()", "_____no_output_____" ] ], [ [ "Some of them are `Series`, some are `Images`, and some are associated `Params` (e.g. covariates). Let's load an `Images` dataset:", "_____no_output_____" ] ], [ [ "images = tsc.loadExample('mouse-images')", "_____no_output_____" ], [ "images", "_____no_output_____" ] ], [ [ "Now every record is an key-value pair where the key is an identifier, and the value is an image", "_____no_output_____" ] ], [ [ "key, value = images.first()", "_____no_output_____" ] ], [ [ "The key is an integer", "_____no_output_____" ] ], [ [ "key", "_____no_output_____" ] ], [ [ "And the value is a two-dimensional array", "_____no_output_____" ] ], [ [ "value.shape", "_____no_output_____" ] ], [ [ "Although `images` is not an array, some syntactic sugar supports easy indexing:", "_____no_output_____" ] ], [ [ "im = images[0]\nimage(im)", "_____no_output_____" ] ], [ [ "And we can now apply simple parallelized image processing routines", "_____no_output_____" ] ], [ [ "im = images.gaussianFilter(3).subsample(3)[0]\nimage(im)", "_____no_output_____" ] ], [ [ "For both `Images` and `Series` data, there are a variety of more complex analyses that can be run on these objects, including massively parallel regression, factorization, registration, feature extraction, and more! See the other tutorials for more information.", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "## Print Cirq Circuit and Statevector", "_____no_output_____" ] ], [ [ "# importing Qiskit\nfrom qiskit import Aer, transpile, assemble\nfrom qiskit import QuantumCircuit, ClassicalRegister, QuantumRegister\nfrom qiskit.visualization import plot_histogram, plot_bloch_multivector\nfrom qiskit.visualization import plot_state_paulivec, plot_state_hinton, plot_state_city\nfrom qiskit.visualization import plot_state_qsphere\n\n# import basic plot tools\nfrom qiskit.visualization import plot_histogram, plot_bloch_multivector\n\nimport matplotlib.pyplot as plt", "_____no_output_____" ], [ "#get backend simulator\nsim = Aer.get_backend('aer_simulator') \n\nqc = QuantumCircuit(3)\nqc.h(0)\nqc.cx(0,1)\nqc.h(2)\nqc.s(2)\n\n#print circuit\nprint(qc)\n\n#draw bloch spheres\nqc.save_statevector()\nstatevector = sim.run(qc).result().get_statevector()\n\nprint(\"\\n\")\nprint(statevector)\nprint(\"\\n\")\n\nplot_bloch_multivector(statevector)", " ┌───┐ \nq_0: ┤ H ├──■──\n └───┘┌─┴─┐\nq_1: ─────┤ X ├\n ┌───┐├───┤\nq_2: ┤ H ├┤ S ├\n └───┘└───┘\n\n\nStatevector([0.5+0.j , 0. +0.j , 0. +0.j , 0.5+0.j , 0. +0.5j, 0. +0.j ,\n 0. +0.j , 0. +0.5j],\n dims=(2, 2, 2))\n\n\n" ], [ "plot_state_city(statevector)", "_____no_output_____" ], [ "plot_state_qsphere(statevector)", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import numpy as np\nimport os\nimport matplotlib.pyplot as plt\nimport cv2\n%matplotlib inline", "_____no_output_____" ], [ "vid = 'D:\\\\Open Field Behavior\\Test 520 - 9 mth - WT - NULL - 3047.mp4'", "_____no_output_____" ], [ "cam = cv2.VideoCapture(vid) \n \ntry: \n \n # creating a folder named data \n if not os.path.exists('data'): \n os.makedirs('data') \n \n# if not created then raise error \nexcept OSError: \n print ('Error: Creating directory of data') \n \n# frame \ncurrentframe = 0\n \nwhile(True): \n \n # reading from frame \n ret,frame = cam.read() \n \n if ret: \n # if video is still left continue creating images \n name = './data/frame' + str(currentframe) + '.jpg'\n print ('Creating...' + name) \n \n # writing the extracted images \n cv2.imwrite(name, frame) \n \n # increasing counter so that it will \n # show how many frames are created \n currentframe += 1\n break\n else: \n break\n \n# Release all space and windows once done \ncam.release() \ncv2.destroyAllWindows() ", "Creating..../data/frame0.jpg\n" ], [ "img = cv2.imread('data/frame0.jpg')\nplt.imshow(img)\nplt.plot(190, 390, 'ro')\nplt.plot(190, 90, 'ro')\nplt.plot(490, 90, 'ro')\nplt.plot(490, 390, 'ro')\nplt.plot((190+490)/2, (390+90)/2, 'bo')", "_____no_output_____" ], [ "center = ((190+490)/2, (390+90)/2)", "_____no_output_____" ], [ "os.listdir('D:\\\\Open field behavior\\plot-poses\\Test 520 - 9 mth - WT - NULL - 3047\\kalman_filter_smoothing_plot_center_back.npy')", "_____no_output_____" ], [ "positions = np.load('D:\\\\Open field behavior\\plot-poses\\Test 523 - 9 mth - HET - NULL - 3063\\kalman_filter_smoothing_plot_center_back.npy')", "_____no_output_____" ], [ "plt.scatter(positions[:, 0], positions[:, 2], .1)\nplt.plot(center[0], center[1], 'ro')\nplt.show()", "_____no_output_____" ], [ "os.listdir('D:\\\\Open field behavior\\plot-poses')", "_____no_output_____" ], [ "import os\nimport matplotlib.pyplot as plt\n%matplotlib inline\nimport numpy as np\n\n\n\n\n\ncenter = ((190+490)/2, (390+90)/2)\nstart = 0; end = 15\nplot_poses = 'D:\\Open field behavior\\plot-poses\\\\'\ndist_from_center = {'WT-NULL': {9: [], 18: []}, 'HET-NULL': {9: [], 18: []}, 'HET-ZFP': {9: [], 18: []}}\nfrom tqdm import tqdm\ndef distance_between_two_points(x1, y1, x2, y2):\n return np.sqrt((x2-x1)**2 + (y2-y1)**2) * (4/31)\n\n\nfor folder in tqdm(os.listdir(plot_poses)):\n mouse_folder = folder\n if 'Test' not in folder:\n continue\n folder = plot_poses + folder\n smoothed_center_back = np.load(folder + '/kalman_filter_smoothing_plot_center_back.npy')\n \n smoothed_cb_x = smoothed_center_back[:,0]\n smoothed_cb_y = smoothed_center_back[:,2]\n \n \n control = mouse_folder.split('mth -')[1][:11].replace(' ', '')\n if control[-1] == '-':\n control = control[:-1]\n month = int(mouse_folder.split('mth')[0][-3:])\n fps = len(smoothed_cb_x) // (15*60)\n \n distances = []\n for i in range(start*60*fps, end*60*fps):\n cb_position = smoothed_center_back[i, [0,2]] # (x,y) of center back\n distances.append(distance_between_two_points(cb_position[0], cb_position[1], center[0], center[1]))\n dist_from_center[control][month].append(np.average(distances))", "100%|████████████████████████████████████████████████████████████████████████████████| 102/102 [00:09<00:00, 10.93it/s]\n" ], [ "import pandas as pd\ndatasets = [dist_from_center]\nnames = [\"Average Distance From Center\"]\n\ntime = str(start) + '-' + str(end)\nfor idx in range(len(datasets)):\n dict_to_df = {'Genotype': [], 'Metric': []}\n dic = datasets[idx]\n print(names[idx])\n for key in dic: \n \n for key2 in dic[key]:\n geno = key + \": \" + str(key2) + \" Month\"\n for dat in dic[key][key2]:\n dict_to_df['Genotype'].append(geno)\n dict_to_df['Metric'].append(dat)\n df = pd.DataFrame.from_dict(dict_to_df)\n print(plot_poses + 'final_csv\\\\' + names[idx] + \"_\" + time + \".csv\")\n df.to_csv(plot_poses + 'final_csv\\\\' + names[idx] + \"_\" + time + \".csv\")", "Average Distance From Center\nD:\\Open field behavior\\plot-poses\\final_csv\\Average Distance From Center_0-15.csv\n" ], [ "center = ((190+490)/2, (390+90)/2)\nstart = 2; end = 5\nplot_poses = 'D:\\Open field behavior\\plot-poses\\\\'\ntime_spent_in_box = {'WT-NULL': {9: [], 18: []}, 'HET-NULL': {9: [], 18: []}, 'HET-ZFP': {9: [], 18: []}}\nfrom tqdm import tqdm\ndef distance_between_two_points(x1, y1, x2, y2):\n return np.sqrt((x2-x1)**2 + (y2-y1)**2) * (4/31)\ncount = 0\n\nfor folder in tqdm(os.listdir(plot_poses)):\n mouse_folder = folder\n if 'Test' not in folder:\n continue\n folder = plot_poses + folder\n smoothed_center_back = np.load(folder + '/kalman_filter_smoothing_plot_center_back.npy')\n \n smoothed_cb_x = smoothed_center_back[:,0]\n smoothed_cb_y = smoothed_center_back[:,2]\n \n \n control = mouse_folder.split('mth -')[1][:11].replace(' ', '')\n if control[-1] == '-':\n control = control[:-1]\n month = int(mouse_folder.split('mth')[0][-3:])\n fps = len(smoothed_cb_x) // (15*60)\n upper_right = (np.array(center) * (4/31) + 15)#* (4/31)\n lower_left = (np.array(center)* (4/31) - 15)#* (4/31)\n lower_right = ((lower_left[0] + 30) , lower_left[1])\n upper_left = ((upper_right[0] - 30), upper_right[1])\n \n plt.scatter(upper_left[0], upper_left[1], label = 'Border', color = 'red')\n plt.scatter(upper_right[0], upper_right[1], color = 'red')\n plt.scatter(lower_right[0], lower_right[1], color = 'red')\n plt.scatter(lower_left[0], lower_left[1], color = 'red')\n plt.plot([upper_left[0], lower_left[0]], [upper_left[1], lower_left[1]], color = 'red')\n plt.plot([upper_left[0], upper_right[0]], [upper_left[1], upper_right[1]], color = 'red')\n plt.plot([lower_left[0], lower_right[0]], [lower_left[1], lower_right[1]], color = 'red')\n plt.plot([upper_right[0], lower_right[0]], [upper_right[1], lower_right[1]], color = 'red')\n \n #plt.scatter(smoothed_cb_x * (4/31), smoothed_cb_y * (4/31), alpha = .01)\n \n time_in_box = 0\n x_bottom = lower_left[1]\n x_top = upper_left[1]\n y_left = lower_left[0]\n y_right = lower_right[0]\n #print('x_bottom: {}, y_right: {}, x_top: {}, y_left:{}'.format(x_bottom, y_left, x_top, y_right))\n label_counter = 2\n labels = []\n print(mouse_folder)\n for i in range(start*60*fps, end*60*fps):\n cb_position = smoothed_center_back[i, [0,2]] *(4/31)# (x,y) of center back\n\n \n in_box = False\n if cb_position[1] < x_top and cb_position[1] > x_bottom:\n if cb_position[0] > y_left and cb_position[0] < y_right:\n in_box = True\n \n #print(in_box)\n color = 'green' if in_box else 'blue'\n if in_box:\n time_in_box += 1\n #label = 'In Box' if in_box else 'Out of Box'\n #labels = list(labels)\n #labels.append(label)\n #label = set(labels)\n #if len(labels) == 2: \n label = ''\n plt.scatter(cb_position[0], cb_position[1] , alpha = .1, color = color)\n \n \n \n time_spent_in_box[control][month].append(time_in_box/fps)\n plt.xlabel('X (cm)')\n plt.ylabel('Y (cm)')\n plt.title('Trajectory Plot Example With 30 x 30 Box')\n plt.savefig('30x30_box.pdf')\n \n plt.show()\n break\n #dist_from_center[control][month].append(np.average(distances))", " 0%| | 0/102 [00:00<?, ?it/s]" ], [ "import pandas as pd\ndatasets = [time_spent_in_box]\nnames = [\"Time Spent In Box 30x30\"]\n\ntime = str(start) + '-' + str(end)\nfor idx in range(len(datasets)):\n dict_to_df = {'Genotype': [], 'Metric': []}\n dic = datasets[idx]\n print(names[idx])\n for key in dic: \n \n for key2 in dic[key]:\n geno = key + \": \" + str(key2) + \" Month\"\n for dat in dic[key][key2]:\n dict_to_df['Genotype'].append(geno)\n dict_to_df['Metric'].append(dat)\n df = pd.DataFrame.from_dict(dict_to_df)\n print(plot_poses + 'final_csv\\\\' + names[idx] + \"_\" + time + \".csv\")\n df.to_csv(plot_poses + 'final_csv\\\\' + names[idx] + \"_\" + time + \".csv\")", "Time Spent In Box 30x30\nD:\\Open field behavior\\plot-poses\\final_csv\\Time Spent In Box 30x30_2-5.csv\n" ], [ "time_spent_in_box", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "from database.strategy import Strategy\nfrom database.sec import SEC\nfrom database.market import Market\nfrom database.weather import Weather\nfrom transformer.date_transformer import DateTransformer\nfrom transformer.column_transformer import ColumnTransformer\nfrom transformer.model_transformer import ModelTransformer\nfrom transformer.product_transformer import ProductTransformer\nfrom transformer.predictor_transformer import PredictorTransformer\nfrom preprocessor.model_preprocessor import ModelPreprocessor\nfrom preprocessor.predictor_preprocessor import PredictorPreprocessor\nfrom modeler.modeler import Modeler as sp\nfrom utils.date_utils import DateUtils\nimport pandas as pd\nimport matplotlib.pyplot as plt\nfrom datetime import datetime, timedelta, timezone\nfrom tqdm import tqdm\nimport math\nimport findspark\nfindspark.init()\nfrom pyspark import SparkContext\nfrom pyspark.sql import SparkSession", "_____no_output_____" ], [ "## Loading Constants\nstart = \"2008-01-01\"\nend = datetime(2021,1,7).strftime(\"%Y-%m-%d\")\n# Loading Databases\nstrat_db = Strategy(\"unity\")\nmarket = Market()\ndataset = \"pdr\"\n## Acquiring Datasets\nmarket.connect()\ntickers = market.retrieve_data(\"sp500\")\nclassification = market.retrieve_data(\"dataset_week_classification\")\nregression = market.retrieve_data(\"dataset_week_regression\")\nfor col in regression.columns:\n if -99999 == regression[col].min():\n regression.drop(col,axis=1,inplace=True)\nfor col in classification.columns:\n if -99999 == classification[col].min():\n classification.drop(col,axis=1,inplace=True)\nmarket.close()\nreload = True", "_____no_output_____" ], [ "gap = 5\nweek_gap = int(gap/5)\ntraining_years = 5", "_____no_output_____" ], [ "## setting up groups\ntickers[\"group\"] = [x % 6 for x in list(tickers.index)]", "_____no_output_____" ], [ "tickers_df = spark.createDataFrame(tickers[[\"Symbol\",\"group\"]])", "_____no_output_____" ], [ "test = spark.createDataFrame(tickers_parallel.value)\n", "_____no_output_____" ], [ "\n# try:\n\n# mr = ModelPreprocessor(ticker)\n# prot = ProductTransformer(ticker,start,end)\n# ticker_regression = all_regression.copy()\n# ticker_regression[\"y\"] = ticker_regression[ticker]\n# ticker_regression[\"y\"] = ticker_regression[\"y\"].shift(-week_gap)\n# ticker_regression = ticker_regression[:-week_gap]\n# ticker_classification = all_classification.copy()\n# ticker_classification[\"y\"] = ticker_classification[ticker]\n# ticker_classification[\"y\"] = ticker_classification[\"y\"].shift(-week_gap)\n# ticker_classification = ticker_classification[:-week_gap]\n# except Exception as e:\n# print(str(e))\n# continue\n# try:\n ## regression_model\n# first = ticker_regression[(ticker_regression[\"year\"] == year - training_years) & (ticker_regression[\"quarter\"] == quarter)].index.values.tolist()[0]\n# last = ticker_regression[(ticker_regression[\"year\"] == year) & (ticker_regression[\"quarter\"] == quarter)].index.values.tolist()[0]\n# rqpd = ticker_regression.iloc[first:last-1]\n# qpd = mr.day_trade_preprocess_regression(rqpd.copy(),ticker,True)\n# rpr = sp.regression(qpd,ranked=False,tf=False,deep=False)\n# ## classification_model\n# first = ticker_classification[(ticker_classification[\"year\"] == year - training_years) & (ticker_classification[\"quarter\"] == quarter)].index.values.tolist()[0]\n# last = ticker_classification[(ticker_classification[\"year\"] == year) & (ticker_classification[\"quarter\"] == quarter)].index.values.tolist()[0]\n# cqpd = ticker_classification.iloc[first:last-1]\n# qpd = mr.day_trade_preprocess_classify(cqpd.copy(),ticker)\n# cpr = sp.classification(qpd,tf=False,deep=False,multioutput=False)\n# price_results = pd.DataFrame([cpr,rpr])\n# product_qpds = []\n# for j in range(len(price_results)):\n# price_result = price_results.iloc[j]\n# if price_result[\"model_type\"] == \"regression\":\n# weekly_price_data = ticker_regression\n# weekly_price_data = weekly_price_data[(weekly_price_data[\"year\"] == year) & (weekly_price_data[\"quarter\"] == quarter)]\n# product_qpd = mr.day_trade_preprocess_regression(weekly_price_data.copy(),ticker,False)\n# else:\n# weekly_price_data = ticker_classification\n# weekly_price_data = weekly_price_data[(weekly_price_data[\"year\"] == year) & (weekly_price_data[\"quarter\"] == quarter)]\n# product_qpd = mr.day_trade_preprocess_classify(weekly_price_data.copy(),ticker)\n# product_qpds.append(product_qpd)\n# sim = prot.merge_weeklies_v2(product_qpds,price_results,year,quarter,\"price\")\n", "_____no_output_____" ], [ "quarterly_range = range(1,2)\nyearly_range = range(2017,2018)\ndef model(group):\n results = []\n all_regression = regression.copy()\n all_classification = classification.copy()\n group_tickers = tickers[tickers[\"group\"]== group]\n model_range = range(group_tickers.index.size)\n for i in tqdm(model_range):\n try:\n ticker = group_tickers.iloc[i][\"Symbol\"].replace(\".\",\"-\")\n mr = ModelPreprocessor(ticker)\n prot = ProductTransformer(ticker,start,end)\n ticker_regression = all_regression.copy()\n ticker_regression[\"y\"] = ticker_regression[ticker]\n ticker_regression[\"y\"] = ticker_regression[\"y\"].shift(-week_gap)\n ticker_regression = ticker_regression[:-week_gap]\n ticker_classification = all_classification.copy()\n ticker_classification[\"y\"] = ticker_classification[ticker]\n ticker_classification[\"y\"] = ticker_classification[\"y\"].shift(-week_gap)\n ticker_classification = ticker_classification[:-week_gap]\n except Exception as e:\n print(str(e))\n continue\n for year in yearly_range:\n for quarter in quarterly_range:\n try:\n ## regression_model\n first = ticker_regression[(ticker_regression[\"year\"] == year - training_years) & (ticker_regression[\"quarter\"] == quarter)].index.values.tolist()[0]\n last = ticker_regression[(ticker_regression[\"year\"] == year) & (ticker_regression[\"quarter\"] == quarter)].index.values.tolist()[0]\n rqpd = ticker_regression.iloc[first:last-1]\n qpd = mr.day_trade_preprocess_regression(rqpd.copy(),ticker,True)\n rpr = sp.regression(qpd,ranked=False,tf=False,deep=False)\n ## classification_model\n first = ticker_classification[(ticker_classification[\"year\"] == year - training_years) & (ticker_classification[\"quarter\"] == quarter)].index.values.tolist()[0]\n last = ticker_classification[(ticker_classification[\"year\"] == year) & (ticker_classification[\"quarter\"] == quarter)].index.values.tolist()[0]\n cqpd = ticker_classification.iloc[first:last-1]\n qpd = mr.day_trade_preprocess_classify(cqpd.copy(),ticker)\n cpr = sp.classification(qpd,tf=False,deep=False,multioutput=False)\n price_results = pd.DataFrame([cpr,rpr])\n product_qpds = []\n for j in range(len(price_results)):\n price_result = price_results.iloc[j]\n if price_result[\"model_type\"] == \"regression\":\n weekly_price_data = ticker_regression\n weekly_price_data = weekly_price_data[(weekly_price_data[\"year\"] == year) & (weekly_price_data[\"quarter\"] == quarter)]\n product_qpd = mr.day_trade_preprocess_regression(weekly_price_data.copy(),ticker,False)\n else:\n weekly_price_data = ticker_classification\n weekly_price_data = weekly_price_data[(weekly_price_data[\"year\"] == year) & (weekly_price_data[\"quarter\"] == quarter)]\n product_qpd = mr.day_trade_preprocess_classify(weekly_price_data.copy(),ticker)\n product_qpds.append(product_qpd)\n sim = prot.merge_weeklies_v2(product_qpds,price_results,year,quarter,\"price\")\n results.append(sim)\n except Exception as e:\n print(ticker,str(e))\n continue\n return pd.con(results)", "_____no_output_____" ], [ "sim = group_parallel.map(lambda x: model(x)).collect()", "_____no_output_____" ], [ "pd.concat(sim)", "_____no_output_____" ], [ "product = pd.concat(sim)\n# product = product.merge(regression,on=[\"year\",\"quarter\",\"week\"],how=\"left\")", "_____no_output_____" ], [ "strat_db.connect()\nstrat_db.store_data(\"parallel_weekly_sim\",product)\nstrat_db.close()", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import pandas as pd\nimport numpy as np\nimport stumpy\nimport sklearn\nimport lightgbm\nimport xgboost\nimport skforecast\nimport statsmodels\nimport matplotlib.pyplot as plt", "_____no_output_____" ], [ "prediction_input = pd.read_parquet(\"data/prediction_input.parquet\")\ninput_dataset_2 = pd.read_parquet(\"data/input_dataset-2.parquet\")", "_____no_output_____" ], [ "plt.plot(prediction_input)\nplt.show()\n", "_____no_output_____" ], [ "df = pd.DataFrame(input_dataset_2)\ndf_prediction = pd.DataFrame(prediction_input)", "_____no_output_____" ], [ "df_prediction.plot()", "_____no_output_____" ], [ "print(df.columns)\ndf.shape\ndf.index\ndf_prediction.head()", "Index(['Unit_4_Power', 'Unit_4_Reactive Power', 'Turbine_Guide Vane Opening',\n 'Turbine_Pressure Drafttube', 'Turbine_Pressure Spiral Casing',\n 'Turbine_Rotational Speed', 'mode', 'Bolt_1_Steel tmp',\n 'Bolt_1_Tensile', 'Bolt_2_Tensile', 'Bolt_3_Tensile', 'Bolt_4_Tensile',\n 'Bolt_5_Tensile', 'Bolt_6_Tensile', 'Bolt_1_Torsion', 'Bolt_2_Torsion',\n 'Bolt_3_Torsion', 'Bolt_4_Torsion', 'Bolt_5_Torsion', 'Bolt_6_Torsion',\n 'lower_bearing_vib_vrt', 'turbine_bearing_vib_vrt'],\n dtype='object')\n" ], [ "figure, ax1 = plt.subplots()\nax1.plot(df.iloc[:, 6])", "_____no_output_____" ], [ "def add_seconds_operational(dataframe):\n # Find index of \"start\" modes in the timeseries\n start_ts = dataframe[dataframe['mode'] == 'start'].index\n # Calculate secods until next \"start\" mode\n secs_since_last_start = (start_ts[1:] - start_ts[:-1]).seconds\n # Extract a list of \"timeslots\" that pairwise indicate a sequence of operation\n last_start_before_counting = start_ts[:-1][secs_since_last_start > 1]\n last_start_before_counting = list(last_start_before_counting)\n last_start_before_counting.append(dataframe.index[-1])\n # Create new feature\n dataframe['sec_since_last_start'] = 0\n for t1, t2 in zip(last_start_before_counting[:-1], last_start_before_counting[1:]):\n dataframe.loc[t1:t2, 'sec_since_last_start'] = range(len(dataframe[t1:t2]))\n # Force \"start\" mde to equal 0 seconds\n dataframe.loc[dataframe['mode'] == 'start', 'sec_since_last_start'] = 0\n # Look at the last timeslot\n dataframe.loc[t1:t2, ['mode', 'sec_since_last_start']]\n\n return dataframe\n\ndf = add_seconds_operational(df)", "_____no_output_____" ], [ "def add_hour_feature(dataframe):\n dataframe['hour'] = dataframe.index.hour\n return dataframe", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "#", "_____no_output_____" ], [ "#", "_____no_output_____" ], [ "print(\"no problem in this cell\")", "no problem in this cell\n" ], [ "print(\"no problem in this cell either\")", "no problem in this cell either\n" ], [ "print(\"always gets stuck in this cell\")", "always gets stuck in this cell\n" ], [ "print(\"even though the execution continues\")", "even though the execution continues\n" ], [ "import time\ntime.sleep(5)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import numpy as np\nfrom keras import callbacks\n\nimport capsulenet", "Using TensorFlow backend.\n/home/watiz/.virtualenvs/siamese_reid/lib/python3.5/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n from ._conv import register_converters as _register_converters\n" ] ], [ [ "## Args", "_____no_output_____" ] ], [ [ "class args:\n save_dir = \"weights/\"\n debug = True\n \n # model\n routings = 1\n \n # hp\n batch_size = 32\n lr = 0.001\n lr_decay = 1.0\n lam_recon = 0.392\n \n # training\n epochs = 3\n shift_fraction = 0.1\n digit = 5", "_____no_output_____" ] ], [ [ "## Load data", "_____no_output_____" ] ], [ [ "(x_train, y_train), (x_test, y_test) = capsulenet.load_mnist()", "_____no_output_____" ] ], [ [ "## Define model", "_____no_output_____" ] ], [ [ "model, eval_model, manipulate_model = capsulenet.CapsNet(input_shape=x_train.shape[1:],\n n_class=len(np.unique(np.argmax(y_train, 1))),\n routings=args.routings)", "_____no_output_____" ] ], [ [ "## Training", "_____no_output_____" ] ], [ [ "capsulenet.train(model=model, data=((x_train, y_train), (x_test, y_test)), args=args)", "INFO:tensorflow:Summary name conv1/kernel:0 is illegal; using conv1/kernel_0 instead.\nINFO:tensorflow:Summary name conv1/bias:0 is illegal; using conv1/bias_0 instead.\nINFO:tensorflow:Summary name primarycap_conv2d/kernel:0 is illegal; using primarycap_conv2d/kernel_0 instead.\nINFO:tensorflow:Summary name primarycap_conv2d/bias:0 is illegal; using primarycap_conv2d/bias_0 instead.\nINFO:tensorflow:Summary name digitcaps/W:0 is illegal; using digitcaps/W_0 instead.\nINFO:tensorflow:Summary name dense_1/kernel:0 is illegal; using dense_1/kernel_0 instead.\nINFO:tensorflow:Summary name dense_1/bias:0 is illegal; using dense_1/bias_0 instead.\nINFO:tensorflow:Summary name dense_2/kernel:0 is illegal; using dense_2/kernel_0 instead.\nINFO:tensorflow:Summary name dense_2/bias:0 is illegal; using dense_2/bias_0 instead.\nINFO:tensorflow:Summary name dense_3/kernel:0 is illegal; using dense_3/kernel_0 instead.\nINFO:tensorflow:Summary name dense_3/bias:0 is illegal; using dense_3/bias_0 instead.\nEpoch 1/3\n187/187 [==============================] - 128s - loss: 0.2853 - capsnet_loss: 0.2520 - decoder_loss: 0.0851 - capsnet_acc: 0.7122 - val_loss: 0.0985 - val_capsnet_loss: 0.0764 - val_decoder_loss: 0.0565 - val_capsnet_acc: 0.9540\nEpoch 2/3\n187/187 [==============================] - 151s - loss: 0.1101 - capsnet_loss: 0.0843 - decoder_loss: 0.0658 - capsnet_acc: 0.9402 - val_loss: 0.0740 - val_capsnet_loss: 0.0522 - val_decoder_loss: 0.0557 - val_capsnet_acc: 0.9740\nEpoch 3/3\n187/187 [==============================] - 153s - loss: 0.0883 - capsnet_loss: 0.0642 - decoder_loss: 0.0615 - capsnet_acc: 0.9545 - val_loss: 0.0647 - val_capsnet_loss: 0.0439 - val_decoder_loss: 0.0530 - val_capsnet_acc: 0.9770\nTrained model saved to 'weights//trained_model.h5'\n" ], [ "capsulenet.test(eval_model, data=(x_test, y_test), args=args)", "------------------------------Begin: test------------------------------\nTest acc: 0.9784\n\nReconstructed images are saved to weights//real_and_recon.png\n------------------------------End: test------------------------------\n" ] ] ]

[ "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "import torch\nimport torch.nn as nn\nimport pandas as pd\nimport numpy as np", "_____no_output_____" ], [ "df = pd.read_csv(\"../data/FullData/full_dialog.csv\").iloc[:, 1:]", "_____no_output_____" ], [ "dialog_list = set(item for item in set(df[\"B2\"]) if item[:3] != \"BAD\")\nall_dialogs = []\n\nfor tag in dialog_list:\n all_dialogs.append(df[df[\"B2\"] == tag])", "_____no_output_____" ], [ "indices = np.arange(len(all_dialogs))\nnp.random.shuffle(indices)\ncut_index = int(len(all_dialogs)*0.9)\ntrain_indices = indices[:cut_index]\nval_indices = indices[cut_index:]", "_____no_output_____" ], [ "raw_train_dialogs = [all_dialogs[i] for i in train_indices]\nraw_val_dialogs = [all_dialogs[i] for i in val_indices]", "_____no_output_____" ], [ "torch.save(raw_train_dialogs, \"raw_train_dialogs.pkl\")\ntorch.save(raw_val_dialogs, \"raw_val_dialogs.pkl\")", "_____no_output_____" ], [ "raw_train_dialogs[0]", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/institutohumai/cursos-python/blob/master/MachineLearning/5_KNNyArbolesDeDecision/KNN_Arboles.ipynb\"> <img src='https://colab.research.google.com/assets/colab-badge.svg' /> </a>\n<div align=\"center\"> Recordá abrir en una nueva pestaña </div>", "_____no_output_____" ], [ "# Modelos no paramétricos: K-Nearest Neighbours y Árboles de decisión\n\nDocumentación:\n- KNN para clasificación: https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsClassifier.html#sklearn.neighbors.KNeighborsClassifier\n- KNN para regresión: https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsRegressor.html#sklearn.neighbors.KNeighborsRegressor\n- Árboles para clasificación: https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html\n- Árboles para regresión: https://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeRegressor.html", "_____no_output_____" ], [ "## Seteo de librerias", "_____no_output_____" ] ], [ [ "import os\nimport matplotlib.pyplot as plt\nfrom matplotlib.colors import ListedColormap\nimport seaborn as sns\nimport pandas as pd\nimport numpy as np\n\nfrom sklearn.datasets import make_classification, make_blobs, load_breast_cancer\nfrom sklearn.preprocessing import OrdinalEncoder\nfrom sklearn.model_selection import train_test_split, GridSearchCV, cross_val_score, KFold\n\n# modelos\nfrom sklearn.neighbors import KNeighborsClassifier\nfrom sklearn.tree import DecisionTreeClassifier, plot_tree\n\nfrom sklearn.metrics import classification_report, confusion_matrix, accuracy_score\n\nDISPLAY_PRECISION = 4\npd.set_option(\"display.precision\", DISPLAY_PRECISION)", "_____no_output_____" ] ], [ [ "# 1. KNN\n\n", "_____no_output_____" ], [ "## 1.1 Introducción: Fronteras de decisión\n\nPara familirizarnos con este modelo y podervisualizar como quedan las fronteras de decisión empezaremos con un problema de clasificación binaria con dos features con un dataset de juguete que generaremos nosotros con la función [make_classification](https://scikit-learn.org/stable/modules/generated/sklearn.datasets.make_classification.html).", "_____no_output_____" ] ], [ [ "# construyamos el dataset para un problema de clasificación binaria de dos dimensiones\nX, y = make_classification(n_samples=200, n_features=2, n_informative=2, n_redundant=0, n_classes=2,n_clusters_per_class=1,\n random_state=1, class_sep=1.1)\n# scatter plot, colores por etiquetas\ndf = pd.DataFrame(dict(x=X[:,0], y=X[:,1], label=y))\ncolors = {0:'red', 1:'blue'}\nfig, ax = plt.subplots()\ngrouped = df.groupby('label')\nfor key, group in grouped:\n group.plot(ax=ax, kind='scatter', x='x', y='y', label=key, color=colors[key])", "_____no_output_____" ], [ "# instanciemos y entrenemos el modelo\nmodel = KNeighborsClassifier(n_neighbors=10,weights='uniform')\nmodel.fit(X,y)", "_____no_output_____" ], [ "# visualicemos las predicciones\n\n# elegimos algunos colores de la lista de colores\ncmap_light = ListedColormap(['#FFAAAA', '#AAFFAA', '#AAAAFF'])\ncmap_bold = ListedColormap(['#FF0000', '#00FF00', '#0000FF'])\n\n# tenemos que armar una grilla y setear un step\nh = .02 \nx_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1\ny_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1\nxx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\nZ = model.predict(np.c_[xx.ravel(), yy.ravel()])\n\n# usamos un pcolormesh\nZ = Z.reshape(xx.shape)\nplt.figure()\nplt.pcolormesh(xx, yy, Z, cmap=cmap_light)\n\n# ploteamos también los puntos de entrenamiento\nplt.scatter(X[:, 0], X[:, 1], c=y, cmap=cmap_bold)\nplt.xlim(xx.min(), xx.max())\nplt.ylim(yy.min(), yy.max())\nplt.title(\"2-Class classification (k = %i, weights = '%s')\"\n % (10, 'uniform'))\n\nplt.show()", "_____no_output_____" ], [ "model = KNeighborsClassifier(n_neighbors=200,weights='uniform')\nmodel.fit(X,y)", "_____no_output_____" ], [ "# visualicemos las predicciones\n\n# elegimos algunos colores de la lista de colores\ncmap_light = ListedColormap(['#FFAAAA', '#AAFFAA', '#AAAAFF'])\ncmap_bold = ListedColormap(['#FF0000', '#00FF00', '#0000FF'])\n\n# tenemos que armar una grilla y setear un step\nh = .02 \nx_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1\ny_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1\nxx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\nZ = model.predict(np.c_[xx.ravel(), yy.ravel()])\n\n# usamos un pcolormesh\nZ = Z.reshape(xx.shape)\nplt.figure()\nplt.pcolormesh(xx, yy, Z, cmap=cmap_light)\n\n# ploteamos también los puntos de entrenamiento\nplt.scatter(X[:, 0], X[:, 1], c=y, cmap=cmap_bold)\nplt.xlim(xx.min(), xx.max())\nplt.ylim(yy.min(), yy.max())\nplt.title(\"2-Class classification (k = %i, weights = '%s')\"\n % (200, 'uniform'))\n\nplt.show()", "_____no_output_____" ], [ "model = KNeighborsClassifier(n_neighbors=200,weights='distance')\nmodel.fit(X,y)", "_____no_output_____" ], [ "# visualicemos las predicciones\n\n# elegimos algunos colores de la lista de colores\ncmap_light = ListedColormap(['#FFAAAA', '#AAFFAA', '#AAAAFF'])\ncmap_bold = ListedColormap(['#FF0000', '#00FF00', '#0000FF'])\n\n# tenemos que armar una grilla y setear un step\nh = .02 \nx_min, x_max = X[:, 0].min() - 1, X[:, 0].max() + 1\ny_min, y_max = X[:, 1].min() - 1, X[:, 1].max() + 1\nxx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\nZ = model.predict(np.c_[xx.ravel(), yy.ravel()])\n\n# usamos un pcolormesh\nZ = Z.reshape(xx.shape)\nplt.figure()\nplt.pcolormesh(xx, yy, Z, cmap=cmap_light)\n\n# ploteamos también los puntos de entrenamiento\nplt.scatter(X[:, 0], X[:, 1], c=y, cmap=cmap_bold)\nplt.xlim(xx.min(), xx.max())\nplt.ylim(yy.min(), yy.max())\nplt.title(\"2-Class classification (k = %i, weights = '%s')\"\n % (200, 'distance'))\n\nplt.show()", "_____no_output_____" ] ], [ [ "## 1.2 Conjunto de datos de cáncer de mama\n\nEl conjunto de datos etiquetado proviene de la \"Base de datos (diagnóstico) de cáncer de mama de Wisconsin\" disponible gratuitamente en la biblioteca sklearn de python. Para obtener más detalles, consulte:\nhttps://archive.ics.uci.edu/ml/datasets/Breast+Cancer+Wisconsin+%28Diagnostic%29\n\nNúmero de muestras: 569\n\nNúmero de funciones: 30 atributos numéricos y predictivos\n\nNúmero de clases: 2\n\nLas características se calculan a partir de una imagen digitalizada de una aspiración con aguja fina (FNA) de una masa mamaria. Describen las características de los núcleos celulares presentes en la imagen. Se calculan diez características de valor real para cada núcleo celular. La media, el error estándar y el \"peor\" o el más grande (la media de los tres valores más grandes) de estas características se calcularon para cada imagen, lo que dio como resultado 30 características. Por ejemplo, las medidas del radio son para el \"radio medio\", el \"error estándar del radio\" y el \"peor radio\". Todos los valores de las características se recodifican con cuatro dígitos significativos.\n\nLas dos clases objetivo corresponden a resultados negativos (benignos) y resultados positivos (malignos).\n\nEste conjunto de datos original se dividirá aleatoriamente en dos conjuntos para fines de entrenamiento y prueba.", "_____no_output_____" ] ], [ [ "data = load_breast_cancer()\n#print(data.DESCR)\n\nprint(\"Descripción:\")\nprint(data.keys()) # dict_keys(['target_names', 'target', 'feature_names', 'data', 'DESCR'])\nprint(\"---\")\n\n# Note that we need to reverse the original '0' and '1' mapping in order to end up with this mapping:\n# Benign = 0 (negative class)\n# Malignant = 1 (positive class)\n\ndata_clases = [data.target_names[1], data.target_names[0]]\ndata_target = [1 if x==0 else 0 for x in list(data.target)]\ndata_features = list(data.feature_names)\n\nprint(\"Clases Target:\")\nprint(\"Clases\", data_clases)\nprint(\"---\")\nprint(\"Distribución de clases n=%d:\" % len(data_target))\nprint(pd.Series(data_target).value_counts())\nprint(\"---\")", "Descripción:\ndict_keys(['data', 'target', 'target_names', 'DESCR', 'feature_names', 'filename'])\n---\nClases Target:\nClases ['benign', 'malignant']\n---\nDistribución de clases n=569:\n0 357\n1 212\ndtype: int64\n---\n" ], [ "pd.DataFrame(data.data[:,:], columns=data_features).info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 569 entries, 0 to 568\nData columns (total 30 columns):\n # Column Non-Null Count Dtype \n--- ------ -------------- ----- \n 0 mean radius 569 non-null float64\n 1 mean texture 569 non-null float64\n 2 mean perimeter 569 non-null float64\n 3 mean area 569 non-null float64\n 4 mean smoothness 569 non-null float64\n 5 mean compactness 569 non-null float64\n 6 mean concavity 569 non-null float64\n 7 mean concave points 569 non-null float64\n 8 mean symmetry 569 non-null float64\n 9 mean fractal dimension 569 non-null float64\n 10 radius error 569 non-null float64\n 11 texture error 569 non-null float64\n 12 perimeter error 569 non-null float64\n 13 area error 569 non-null float64\n 14 smoothness error 569 non-null float64\n 15 compactness error 569 non-null float64\n 16 concavity error 569 non-null float64\n 17 concave points error 569 non-null float64\n 18 symmetry error 569 non-null float64\n 19 fractal dimension error 569 non-null float64\n 20 worst radius 569 non-null float64\n 21 worst texture 569 non-null float64\n 22 worst perimeter 569 non-null float64\n 23 worst area 569 non-null float64\n 24 worst smoothness 569 non-null float64\n 25 worst compactness 569 non-null float64\n 26 worst concavity 569 non-null float64\n 27 worst concave points 569 non-null float64\n 28 worst symmetry 569 non-null float64\n 29 worst fractal dimension 569 non-null float64\ndtypes: float64(30)\nmemory usage: 133.5 KB\n" ], [ "# separamos un 25% para test/held-out\nX = pd.DataFrame(data.data[:,:], columns=data_features)\ny = pd.Series(data_target)\n\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=0)", "_____no_output_____" ] ], [ [ "## 1.3 Overfitting: cantidad de vecinos y pesos", "_____no_output_____" ] ], [ [ "# veamos como le va a nuestro modelo variando la cantidad de vecinos y el tipo de peso\nvalores_k = list(range(1,50,4))\nresultados_train_u = []\nresultados_test_u = []\nresultados_train_w = []\nresultados_test_w = []\n\nfor k in valores_k:\n # instanciamos el modelo uniforme\n clf_u = KNeighborsClassifier(n_neighbors=k, weights='uniform')\n clf_u.fit(X_train, y_train)\n y_train_pred = clf_u.predict(X_train)\n y_pred = clf_u.predict(X_test)\n resultados_train_u.append(accuracy_score(y_train, y_train_pred))\n resultados_test_u.append(accuracy_score(y_test, y_pred))\n\n clf_w = KNeighborsClassifier(n_neighbors=k, weights='distance')\n clf_w.fit(X_train, y_train)\n y_train_pred = clf_w.predict(X_train)\n y_pred = clf_w.predict(X_test)\n resultados_train_w.append(accuracy_score(y_train, y_train_pred))\n resultados_test_w.append(accuracy_score(y_test, y_pred))", "_____no_output_____" ], [ "# veamos que paso en cada caso\nf, ax = plt.subplots(1,2,figsize=(14,5),sharey=True)\nax[0].plot(valores_k, resultados_train_u, valores_k, resultados_test_u);\nax[0].legend(['pesos uniformes - train', 'pesos uniformes - test']);\nax[0].set(xlabel='k',ylabel='accuracy');\n\nax[1].plot(valores_k, resultados_train_w, valores_k, resultados_test_w);\nax[1].legend(['pesos distancia - train', 'pesos distancia - test']);\nax[1].set(xlabel='k');", "_____no_output_____" ], [ "# ahora busquemos nuestro mejor modelo usando validacion cruzada y gridsearchcv pero incluyamos otra distancia!\nmodel = KNeighborsClassifier()\nn_neighbors = np.array([1,2,3,5,8,10,15,20,30,50])\nparam_grid = {'n_neighbors': n_neighbors, \n 'weights':['uniform', 'distance'], \n 'metric':['euclidean', 'chebyshev', 'manhattan']}\ngrid = GridSearchCV(estimator=model, param_grid=param_grid)\ngrid.fit(X_train, y_train)\nprint(grid.best_params_)\npd.DataFrame(grid.cv_results_).sample(3)", "{'metric': 'manhattan', 'n_neighbors': 10, 'weights': 'distance'}\n" ], [ "# evaluemos este clasificador usando el classification report\nprint(classification_report(y_test, grid.best_estimator_.predict(X_test), target_names=data_clases))", " precision recall f1-score support\n\n benign 0.96 0.98 0.97 90\n malignant 0.96 0.92 0.94 53\n\n accuracy 0.96 143\n macro avg 0.96 0.95 0.95 143\nweighted avg 0.96 0.96 0.96 143\n\n" ] ], [ [ "## 1.4 Efectos de escala\n\nDado que KNN esta basado en distancias si no usamos una distancia que involucra la varianza entre variables como la distancia de Mahalabois, nuestro modelo se verá afectado\n\n\n\n\n\nhttps://stats.stackexchange.com/questions/287425/why-do-you-need-to-scale-data-in-knn", "_____no_output_____" ] ], [ [ "XX,yy = make_classification(n_samples=400,n_features=2,n_classes=2,\n n_redundant=0,n_informative=2,\n n_clusters_per_class=2,random_state=48)\n\nXX[:,0] = XX[:,0]*30 + 150\n\nprint('Media x: {}'.format(np.mean(XX[:,0])))\nprint('SD x: {}'.format(np.std(XX[:,0])))\n\nprint('Media y: {}'.format(np.mean(XX[:,1])))\nprint('SD y: {}'.format(np.std(XX[:,1])))", "Media x: 150.4912351511699\nSD x: 39.90933364722069\nMedia y: 0.011031649918614495\nSD y: 1.2066153568418274\n" ], [ "kf = KFold(n_splits=5)\nknn = KNeighborsClassifier(n_neighbors=5)\nknn.fit(XX, yy)\nprint(cross_val_score(knn, XX, yy, cv=kf).mean())", "0.885\n" ], [ "from sklearn.preprocessing import StandardScaler\n\nscaler = StandardScaler()\nXX_scaled = scaler.fit_transform(XX)\nprint('Media x: {}'.format(np.mean(XX_scaled[:,0])))\nprint('SD x: {}'.format(np.std(XX_scaled[:,0])))\n\nprint('Media y: {}'.format(np.mean(XX_scaled[:,1])))\nprint('SD y: {}'.format(np.std(XX_scaled[:,1])))", "Media x: 1.4210854715202004e-16\nSD x: 1.0000000000000004\nMedia y: -1.1102230246251566e-17\nSD y: 0.9999999999999997\n" ], [ "knn = KNeighborsClassifier(n_neighbors=5)\nknn.fit(XX, yy)\nprint(cross_val_score(knn, XX_scaled, yy, cv=kf).mean())", "0.9675\n" ] ], [ [ "***\n# 2. Árboles de decisión\n\nContinuaremos trabajando con el dataset de cancer de mama para familiarizarnos con los árboles de decisión", "_____no_output_____" ], [ "## 2.1 Mi primer arbolito", "_____no_output_____" ] ], [ [ "# instanciemos el modelo y entremoslo en el conjunto de autos\narbol = DecisionTreeClassifier(criterion='gini', max_depth=2, min_samples_leaf=1, min_samples_split=2, ccp_alpha=0)\narbol.fit(X_train,y_train)\naccuracy_score(y_train, arbol.predict(X_train))", "_____no_output_____" ], [ "# veamos que tan bien le fue a este modelo\nprint(classification_report(y_true=y_test,y_pred=arbol.predict(X_test)))", " precision recall f1-score support\n\n 0 0.93 0.94 0.94 90\n 1 0.90 0.89 0.90 53\n\n accuracy 0.92 143\n macro avg 0.92 0.92 0.92 143\nweighted avg 0.92 0.92 0.92 143\n\n" ], [ "# visualicemos los errores de este árbol en una matriz de confusión\ncf_matrix = confusion_matrix(y_test, arbol.predict(X_test))\nsns.heatmap(cf_matrix, annot=True);", "_____no_output_____" ] ], [ [ "## 2.2 Feature importance\n\nLos árboles nos permiten definir una manera de medir la importancia de los features (o *Feature Importances*) basado en la ganancia de información obtenida cada vez que se utilizo cada feature para hacer un split. Para esto, una vez entrando el árbol, el método que utilizaremos es: \n\n```\n# arbol.feature_importances_\n```\n\n", "_____no_output_____" ] ], [ [ "# calculando las 5 feature importances mas altas\nimportances = pd.Series(arbol.feature_importances_).sort_values(ascending=False)[:5]\nimportances", "_____no_output_____" ], [ "f5_names = list(pd.Series(data.feature_names)[importances.index.to_list()])\nfig, ax = plt.subplots()\nimportances.plot.barh(ax=ax)\nax.set_yticklabels(f5_names)\nax.invert_yaxis()", "_____no_output_____" ] ], [ [ "## 2.3 Desbalance de clases\nComo este dataset tiene un desbalance de clases, podes incluir eso en el modelo utilizando el parámetro class_weight que nos permite manejar directamente el desbalance\n", "_____no_output_____" ] ], [ [ "arbol = DecisionTreeClassifier(criterion='gini', max_depth=2, min_samples_leaf=1,\n min_samples_split=2, ccp_alpha=0, class_weight=\"balanced\")\narbol.fit(X_train, y_train)\naccuracy_score(y_train, arbol.predict(X_train))", "_____no_output_____" ], [ "print(classification_report(y_true=y_test,y_pred=arbol.predict(X_test)))", " precision recall f1-score support\n\n 0 1.00 0.90 0.95 90\n 1 0.85 1.00 0.92 53\n\n accuracy 0.94 143\n macro avg 0.93 0.95 0.93 143\nweighted avg 0.95 0.94 0.94 143\n\n" ], [ "cf_matrix = confusion_matrix(y_test, arbol.predict(X_test))\nsns.heatmap(cf_matrix, annot=True);", "_____no_output_____" ] ], [ [ "## 2.4 Visualización\n\nPara visualizar el árbol sklearn tiene el método tree.plot_tree:", "_____no_output_____" ] ], [ [ "plot_tree(arbol);", "_____no_output_____" ] ], [ [ "Podemos obtener una representación mas estilizada con la ayuda de las librerías *graphviz* + *dot*. Ref: https://towardsdatascience.com/visualizing-decision-trees-with-python-scikit-learn-graphviz-matplotlib-1c50b4aa68dc", "_____no_output_____" ] ], [ [ "# libreria\nfrom sklearn.externals.six import StringIO \nfrom IPython.display import Image \nfrom sklearn.tree import export_graphviz\nimport pydotplus\nimport matplotlib.pyplot as plt\n\ndot_data = StringIO()\nexport_graphviz(arbol, out_file=dot_data, \n filled=True, rounded=True,\n special_characters=True)\ngraph = pydotplus.graph_from_dot_data(dot_data.getvalue()) \nImage(graph.create_png())", "/usr/local/lib/python3.7/dist-packages/sklearn/externals/six.py:31: FutureWarning: The module is deprecated in version 0.21 and will be removed in version 0.23 since we've dropped support for Python 2.7. Please rely on the official version of six (https://pypi.org/project/six/).\n \"(https://pypi.org/project/six/).\", FutureWarning)\n" ] ], [ [ "## 2.5 Overfitting: profundidad del árbol y post-pruning\n\nDado que los árboles son modelos que tienden a overfittear tenemos que recurrir a distintas técnicas para mitigar este problema. Veamos primero el efecto de la profundidad del árbol en el trade-off sesgo varianza.", "_____no_output_____" ] ], [ [ "profundidad = list(range(1,20))\nresultados_train = []\nresultados_test = []\n\nfor depth in profundidad:\n # instanciamos el modelo uniforme\n arbol = DecisionTreeClassifier(criterion='gini', max_depth=depth, min_samples_leaf=1, min_samples_split=2, ccp_alpha=0, class_weight=\"balanced\")\n arbol.fit(X_train, y_train)\n y_train_pred = arbol.predict(X_train)\n y_pred = arbol.predict(X_test)\n resultados_train.append(accuracy_score(y_train, y_train_pred))\n resultados_test.append(accuracy_score(y_test, y_pred))", "_____no_output_____" ], [ "# veamos que paso en cada caso\nf, ax = plt.subplots(1,1,figsize=(14,5),sharey=True)\nax.plot(profundidad, resultados_train, profundidad, resultados_test);\nax.legend(['accuracy train', 'accuracy test']);\nax.set(xlabel='profundidad',ylabel='accuracy');", "_____no_output_____" ], [ "# veamos que pasa con un árbol sin corte de profundidad\nnp.random.seed(2021)\narbol = DecisionTreeClassifier(criterion='gini', ccp_alpha=0)\narbol.fit(X_train, y_train)\n#print(classification_report(y_true=y_test,y_pred=arbol.predict(X_test)))\nprint('Accuracy en entrenamiento: %f' % accuracy_score(y_train,arbol.predict(X_train)))\nprint('Accuracy en test: %f' % accuracy_score(y_test,arbol.predict(X_test)))", "Accuracy en entrenamiento: 1.000000\nAccuracy en test: 0.909091\n" ], [ "# grafiquemos este árbol\ndot_data = StringIO()\nexport_graphviz(arbol, out_file=dot_data, \n filled=True, rounded=True,\n special_characters=True)\ngraph = pydotplus.graph_from_dot_data(dot_data.getvalue()) \nImage(graph.create_png())", "_____no_output_____" ] ], [ [ "Una técnica que nos permite mitigar el overfitting es lo que se conoce como post-prunning. El objetivo de esta técnica es *podar* el árbol entrenado, penalizando de alguna forma los árboles más complejos. El algortimo de poda que tenemos implementado en Scikit-Learn es el [Minimal Cost-Complexity Pruning](https://scikit-learn.org/stable/modules/tree.html#minimal-cost-complexity-pruning). El hiperparámetro que controla esta penalización es ccp_alpha$\\geq 0$, cuando este hiperparámetro es 0, no realizamos ningún tipo de poda, y a medida que aumentamos dicho hiperparámetro penalizaremos más fuertemente la cantidad de nodos terminales del árbol.", "_____no_output_____" ] ], [ [ "arbol = DecisionTreeClassifier(criterion='gini', ccp_alpha=0.01)\narbol.fit(X_train, y_train)\n#print(classification_report(y_true=y_test,y_pred=arbol.predict(X_test)))\nprint('Accuracy en entrenamiento: %f' % accuracy_score(y_train,arbol.predict(X_train)))\nprint('Accuracy en test: %f' % accuracy_score(y_test,arbol.predict(X_test)))", "Accuracy en entrenamiento: 0.971831\nAccuracy en test: 0.916084\n" ], [ "dot_data = StringIO()\nexport_graphviz(arbol, out_file=dot_data, \n filled=True, rounded=True,\n special_characters=True)\ngraph = pydotplus.graph_from_dot_data(dot_data.getvalue()) \nImage(graph.create_png())", "_____no_output_____" ], [ "# veamos como afecta el rendimiento y la profundidad del árbol\nccp_alpha_vals = np.arange(0,1,0.05)\nresultados_train = []\nresultados_test = []\nprofundidad = []\n\nfor ccp in ccp_alpha_vals:\n # instanciamos el modelo uniforme\n arbol = DecisionTreeClassifier(criterion='gini', ccp_alpha=ccp)\n arbol.fit(X_train, y_train)\n # guardamos la profundidad del árbol\n profundidad.append(arbol.tree_.max_depth)\n y_train_pred = arbol.predict(X_train)\n y_pred = arbol.predict(X_test)\n resultados_train.append(accuracy_score(y_train, y_train_pred))\n resultados_test.append(accuracy_score(y_test, y_pred))", "_____no_output_____" ], [ "f,ax = plt.subplots(2,1,figsize=(12,8),sharex=True)\nax[0].plot(ccp_alpha_vals, resultados_train, ccp_alpha_vals, resultados_test);\nax[0].legend(['accuracy train', 'accuracy test']);\nax[0].set(xlabel='ccp_alpha',ylabel='Accuracy');\nax[1].plot(ccp_alpha_vals, profundidad)\nax[1].set(xlabel='ccp_alpha',ylabel='Profundidad');", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# import necessary packages\n\nimport tensorflow as tf\nimport pandas as pd", "_____no_output_____" ], [ "# read vulnerability data\n\ndrupal = pd.read_csv('datasets/vulnerability/drupal.csv')\nmoodle = pd.read_csv('datasets/vulnerability/moodle.csv')\nphpmyadmin = pd.read_csv('datasets/vulnerability/phpmyadmin.csv')", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code" ] ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ "BSD-3-Clause" ]

[ [ [ "### This notebook trains a N2V network in the first step and then finetunes it for segmentation.", "_____no_output_____" ] ], [ [ "# We import all our dependencies.\nimport warnings\nwarnings.filterwarnings('ignore')\nimport sys\nsys.path.append('../../')\nfrom voidseg.models import Seg, SegConfig\nfrom n2v.models import N2VConfig, N2V\nimport numpy as np\nfrom csbdeep.utils import plot_history\nfrom voidseg.utils.misc_utils import combine_train_test_data, shuffle_train_data, augment_data\nfrom voidseg.utils.seg_utils import *\nfrom n2v.utils.n2v_utils import manipulate_val_data\nfrom voidseg.utils.compute_precision_threshold import compute_threshold, precision\nfrom keras.optimizers import Adam\nfrom matplotlib import pyplot as plt\nfrom scipy import ndimage\nimport tensorflow as tf\nimport keras.backend as K\nimport urllib\nimport os\nimport zipfile\nfrom tqdm import tqdm, tqdm_notebook", "Using TensorFlow backend.\n" ] ], [ [ "### Download DSB2018 data.\n\nFrom the Kaggle 2018 Data Science Bowl challenge, we take the same subset of data as has been used [here](https://github.com/mpicbg-csbd/stardist), showing a diverse collection of cell nuclei imaged by various fluorescence microscopes. We extracted 4870 image patches of size 128×128 from the training set and added Gaussian noise with mean 0 and sigma = 10 (n10), 20 (n20) and 40 (n40). This notebook shows results for n40 images.", "_____no_output_____" ] ], [ [ "# create a folder for our data\nif not os.path.isdir('./data'):\n os.mkdir('data')\n\n# check if data has been downloaded already\nzipPath=\"data/DSB.zip\"\nif not os.path.exists(zipPath):\n #download and unzip data\n data = urllib.request.urlretrieve('https://owncloud.mpi-cbg.de/index.php/s/LIN4L4R9b2gebDX/download', zipPath)\n with zipfile.ZipFile(zipPath, 'r') as zip_ref:\n zip_ref.extractall(\"data\")", "_____no_output_____" ] ], [ [ "The downloaded data is in `npz` format and the cell below extracts the training, validation and test data as numpy arrays", "_____no_output_____" ] ], [ [ "trainval_data = np.load('data/DSB/train_data/dsb2018_TrainVal40.npz')\ntest_data = np.load('data/DSB/test_data/dsb2018_Test40.npz', allow_pickle=True)\ntrain_images = trainval_data['X_train']\nval_images = trainval_data['X_val']\ntest_images = test_data['X_test']\n\ntrain_masks = trainval_data['Y_train']\nval_masks = trainval_data['Y_val']\ntest_masks = test_data['Y_test']", "_____no_output_____" ], [ "print(\"Shape of train_images: \", train_images.shape, \", Shape of train_masks: \", train_masks.shape)\nprint(\"Shape of val_images: \", val_images.shape, \", Shape of val_masks: \", val_masks.shape)\nprint(\"Shape of test_images: \", test_images.shape, \", Shape of test_masks: \", test_masks.shape)", "Shape of train_images: (3800, 128, 128) , Shape of train_masks: (3800, 128, 128)\nShape of val_images: (670, 128, 128) , Shape of val_masks: (670, 128, 128)\nShape of test_images: (50,) , Shape of test_masks: (50,)\n" ] ], [ [ "### Data preparation for training a N2V network\n\nSince, we can use all the noisy data for training N2V network, we combine the noisy train_images and test_images and use them as input to the N2V network.", "_____no_output_____" ] ], [ [ "X, Y = combine_train_test_data(X_train=train_images,Y_train=train_masks,X_test=test_images,Y_test=test_masks)\nprint(\"Combined Dataset Shape\", X.shape)", "Combined Dataset Shape (4300, 128, 128)\n" ], [ "X_val = val_images\nY_val = val_masks", "_____no_output_____" ] ], [ [ "Next, we shuffle the training pairs and augment the training and validation data.", "_____no_output_____" ] ], [ [ "random_seed = 1 # Seed to shuffle training data (annotated GT and raw image pairs)\n\nX, Y = shuffle_train_data(X, Y, random_seed = random_seed)\nprint(\"Training Data \\n..................\")\nX, Y = augment_data(X, Y)\nprint(\"\\n\")\nprint(\"Validation Data \\n..................\")\nX_val, Y_val = augment_data(X_val, Y_val)", "Training Data \n..................\nRaw image size after augmentation (34400, 128, 128)\nMask size after augmentation (34400, 128, 128)\n\n\nValidation Data \n..................\nRaw image size after augmentation (5360, 128, 128)\nMask size after augmentation (5360, 128, 128)\n" ], [ "# Adding channel dimension\nX = X[..., np.newaxis]\nprint(X.shape)\nX_val = X_val[..., np.newaxis]\nprint(X_val.shape)", "(34400, 128, 128, 1)\n(5360, 128, 128, 1)\n" ] ], [ [ "Let's look at one of our training and validation patches.", "_____no_output_____" ] ], [ [ "sl=0\nplt.figure(figsize=(14,7))\nplt.subplot(1,2,1)\nplt.imshow(X[sl,...,0], cmap='gray')\nplt.title('Training Patch');\nplt.subplot(1,2,2)\nplt.imshow(X_val[sl,...,0], cmap='gray')\nplt.title('Validation Patch');", "_____no_output_____" ] ], [ [ "### Configure N2V Network", "_____no_output_____" ], [ "The data preparation for training a denoising N2V network is now done. Next, we configure N2V network by specifying `N2VConfig` parameters.", "_____no_output_____" ] ], [ [ "config = N2VConfig(X, unet_kern_size=3, n_channel_out=1,train_steps_per_epoch=400, train_epochs=200, \n train_loss='mse', batch_norm=True, \n train_batch_size=128, n2v_perc_pix=0.784, n2v_patch_shape=(64, 64),\n unet_n_first = 32,\n unet_residual = False,\n n2v_manipulator='uniform_withCP', n2v_neighborhood_radius=5, unet_n_depth=4)\n\n# Let's look at the parameters stored in the config-object.\nvars(config)", "_____no_output_____" ], [ "# a name used to identify the model\nmodel_name = 'n40_denoising'\n# the base directory in which our model will live\nbasedir = 'models'\n# We are now creating our network model.\nmodel = N2V(config, model_name, basedir=basedir)\nmodel.prepare_for_training(metrics=())", "_____no_output_____" ] ], [ [ "Now, we begin training the denoising N2V model. In case, a trained model is available, that model is loaded else a new model is trained.", "_____no_output_____" ] ], [ [ "# We are ready to start training now.\nquery_weightpath = os.getcwd()+\"/models/\"+model_name\nweights_present = False\nfor file in os.listdir(query_weightpath):\n if(file == \"weights_best.h5\"):\n print(\"Found weights of a trained N2V network, loading it for prediction!\")\n weights_present = True \n break \nif(weights_present):\n model.load_weights(\"weights_best.h5\")\nelse:\n print(\"Did not find weights of a trained N2V network, training one from scratch!\")\n history = model.train(X, X_val)", "Found weights of a trained N2V network, loading it for prediction!\n" ] ], [ [ "### Data preparation for segmentation step\n\nNext, we normalize all raw data with the mean and std (standard deviation) of the raw `train_images`. Then, we shuffle the raw training images and the correponding Ground Truth (GT). Lastly, we fractionate the training pairs of raw images and corresponding GT to realize the case where not enough annotated, training data is available. For this fractionation, please specify `fraction` parameter below. It should be between 0 (exclusive) and 100 (inclusive).", "_____no_output_____" ] ], [ [ "fraction = 2 # Fraction of annotated GT and raw image pairs to use during training. \nrandom_seed = 1 # Seed to shuffle training data (annotated GT and raw image pairs).\n\n\nassert 0 <fraction<= 100, \"Fraction should be between 0 and 100\"\nmean, std = np.mean(train_images), np.std(train_images)\n\nX_normalized = normalize(train_images, mean, std)\nX_val_normalized = normalize(val_images, mean, std)\nX_test_normalized = normalize(test_images, mean, std)\n\nX_shuffled, Y_shuffled = shuffle_train_data(X_normalized, train_masks, random_seed = random_seed)\nX_frac, Y_frac = fractionate_train_data(X_shuffled, Y_shuffled, fraction = fraction)\nprint(\"Training Data \\n..................\")\nX, Y_train_masks = augment_data(X_frac, Y_frac)\nprint(\"\\n\")\nprint(\"Validation Data \\n..................\")\nX_val, Y_val_masks = augment_data(X_val_normalized, val_masks)", "Training Data \n..................\nRaw image size after augmentation (608, 128, 128)\nMask size after augmentation (608, 128, 128)\n\n\nValidation Data \n..................\nRaw image size after augmentation (5360, 128, 128)\nMask size after augmentation (5360, 128, 128)\n" ] ], [ [ "Next, we do a one-hot encoding of training and validation labels for training a 3-class U-Net. One-hot encoding will extract three channels from each labelled image, where the channels correspond to background, foreground and border.", "_____no_output_____" ] ], [ [ "X = X[...,np.newaxis]\nY = convert_to_oneHot(Y_train_masks)\nX_val = X_val[...,np.newaxis]\nY_val = convert_to_oneHot(Y_val_masks)\nprint(X.shape, Y.shape)\nprint(X_val.shape, Y_val.shape)", "(608, 128, 128, 1) (608, 128, 128, 3)\n(5360, 128, 128, 1) (5360, 128, 128, 3)\n" ] ], [ [ "Let's look at one of our validation patches.", "_____no_output_____" ] ], [ [ "sl=0\nplt.figure(figsize=(20,5))\nplt.subplot(1,4,1)\nplt.imshow(X_val[sl,...,0])\nplt.title('Raw validation image')\nplt.subplot(1,4,2)\nplt.imshow(Y_val[sl,...,0])\nplt.title('1-hot encoded background')\nplt.subplot(1,4,3)\nplt.imshow(Y_val[sl,...,1])\nplt.title('1-hot encoded foreground')\nplt.subplot(1,4,4)\nplt.imshow(Y_val[sl,...,2])\nplt.title('1-hot encoded border')", "_____no_output_____" ] ], [ [ "### Configure Segmentation Network\n\nThe data preparation for segmentation is now done. Next, we configure a segmentation network by specifying `SegConfig` parameters. For example, one can increase `train_epochs` to get even better results at the expense of a longer computation. (This holds usually true for a large `fraction`.)", "_____no_output_____" ] ], [ [ "relative_weights = [1.0,1.0,5.0] # Relative weight of background, foreground and border class for training\n\nconfig = SegConfig(X, unet_kern_size=3, relative_weights = relative_weights,\n train_steps_per_epoch=400, train_epochs=3, batch_norm=True, \n train_batch_size=128, unet_n_first = 32, unet_n_depth=4)\n\n# Let's look at the parameters stored in the config-object.\n# a name used to identify the model\nmodel_name = 'seg_finetune'\n# the base directory in which our model will live\nbasedir = 'models'\n# We are now creating our network model.\nseg_model = Seg(config, model_name, basedir=basedir)\nvars(config)", "_____no_output_____" ] ], [ [ "##### For finetuning, we initialize segmentation network with the best weights of the denoising N2V network trained above.", "_____no_output_____" ] ], [ [ "ft_layers = seg_model.keras_model.layers\nn2v_layers = model.keras_model.layers\n\nfor i in range(0, len(n2v_layers)-2):\n ft_layers[i].set_weights(n2v_layers[i].get_weights())\n\nfor l in seg_model.keras_model.layers:\n l.trainable=True", "_____no_output_____" ] ], [ [ "Now, we begin training the model for segmentation.", "_____no_output_____" ] ], [ [ "seg_model.train(X, Y, (X_val, Y_val))", "Epoch 1/3\n400/400 [==============================] - 152s 380ms/step - loss: 0.3040 - seg_crossentropy: 0.3040 - val_loss: 0.2867 - val_seg_crossentropy: 0.2867\nEpoch 2/3\n400/400 [==============================] - 143s 358ms/step - loss: 0.1202 - seg_crossentropy: 0.1202 - val_loss: 0.3895 - val_seg_crossentropy: 0.3895\nEpoch 3/3\n400/400 [==============================] - 142s 354ms/step - loss: 0.0760 - seg_crossentropy: 0.0760 - val_loss: 0.4506 - val_seg_crossentropy: 0.4506\n\nLoading network weights from 'weights_best.h5'.\n" ] ], [ [ "### Computing the best threshold on validation images (to maximize Average Precision score). The threshold so obtained will be used to get hard masks from probability images to be predicted on test images.", "_____no_output_____" ] ], [ [ "threshold=seg_model.optimize_thresholds(X_val_normalized.astype(np.float32), val_masks)", "Computing best threshold: \n" ] ], [ [ "### Prediction on test images to get segmentation result", "_____no_output_____" ] ], [ [ "predicted_images, precision_result=seg_model.predict_label_masks(X_test_normalized, test_masks, threshold)\nprint(\"Average precision over all test images at IOU = 0.5: \", precision_result)", "Average precision over all test images at IOU = 0.5: 0.6144172916479654\n" ], [ "plt.figure(figsize=(10,10))\nplt.subplot(1,2,1)\nplt.imshow(predicted_images[22])\nplt.title('Prediction')\nplt.subplot(1,2,2)\nplt.imshow(test_masks[22])\nplt.title('Ground Truth')", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/txusser/Master_IA_Sanidad/blob/main/2.3.1_3_Maneras_de_Programar_a_una_Red_Neuronal.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ], [ "# 3 Maneras de Programar a una Red Neuronal - DOTCSV\n\n## Código inicial", "_____no_output_____" ] ], [ [ "import numpy as np\nimport scipy as sc\nimport matplotlib.pyplot as plt\n\nfrom sklearn.datasets import make_circles\n\n# Creamos nuestros datos artificiales, donde buscaremos clasificar \n# dos anillos concéntricos de datos. \nX, Y = make_circles(n_samples=500, factor=0.5, noise=0.05)\n\n# Resolución del mapa de predicción.\nres = 100 \n\n# Coordendadas del mapa de predicción.\n_x0 = np.linspace(-1.5, 1.5, res)\n_x1 = np.linspace(-1.5, 1.5, res)\n\n# Input con cada combo de coordenadas del mapa de predicción.\n_pX = np.array(np.meshgrid(_x0, _x1)).T.reshape(-1, 2)\n\n# Objeto vacio a 0.5 del mapa de predicción.\n_pY = np.zeros((res, res)) + 0.5\n\n# Visualización del mapa de predicción.\nplt.figure(figsize=(8, 8))\nplt.pcolormesh(_x0, _x1, _pY, cmap=\"coolwarm\", vmin=0, vmax=1)\n\n# Visualización de la nube de datos.\nplt.scatter(X[Y == 0,0], X[Y == 0,1], c=\"skyblue\")\nplt.scatter(X[Y == 1,0], X[Y == 1,1], c=\"salmon\")\n\nplt.tick_params(labelbottom=False, labelleft=False)", "_____no_output_____" ] ], [ [ "## Tensorflow", "_____no_output_____" ] ], [ [ "import tensorflow as tf\n\nfrom matplotlib import animation\nfrom IPython.core.display import display, HTML\n\n# Definimos los puntos de entrada de la red, para la matriz X e Y.\niX = tf.placeholder('float', shape=[None, X.shape[1]])\niY = tf.placeholder('float', shape=[None])\n\nlr = 0.01 # learning rate\nnn = [2, 16, 8, 1] # número de neuronas por capa.\n\n# Capa 1\nW1 = tf.Variable(tf.random_normal([nn[0], nn[1]]), name='Weights_1')\nb1 = tf.Variable(tf.random_normal([nn[1]]), name='bias_1')\n\nl1 = tf.nn.relu(tf.add(tf.matmul(iX, W1), b1))\n\n# Capa 2\nW2 = tf.Variable(tf.random_normal([nn[1], nn[2]]), name='Weights_2')\nb2 = tf.Variable(tf.random_normal([nn[2]]), name='bias_2')\n\nl2 = tf.nn.relu(tf.add(tf.matmul(l1, W2), b2))\n\n# Capa 3\nW3 = tf.Variable(tf.random_normal([nn[2], nn[3]]), name='Weights_3')\nb3 = tf.Variable(tf.random_normal([nn[3]]), name='bias_3')\n\n# Vector de predicciones de Y.\npY = tf.nn.sigmoid(tf.add(tf.matmul(l2, W3), b3))[:, 0]\n\n\n# Evaluación de las predicciones.\nloss = tf.losses.mean_squared_error(pY, iY)\n\n# Definimos al optimizador de la red, para que minimice el error.\noptimizer = tf.train.GradientDescentOptimizer(learning_rate=0.05).minimize(loss)\n\nn_steps = 1000 # Número de ciclos de entrenamiento.\n\niPY = [] # Aquí guardaremos la evolución de las predicción, para la animación.\n\nwith tf.Session() as sess:\n \n # Inicializamos todos los parámetros de la red, las matrices W y b.\n sess.run(tf.global_variables_initializer())\n \n # Iteramos n pases de entrenamiento.\n for step in range(n_steps):\n \n # Evaluamos al optimizador, a la función de coste y al tensor de salida pY. \n # La evaluación del optimizer producirá el entrenamiento de la red.\n _, _loss, _pY = sess.run([optimizer, loss, pY], feed_dict={ iX : X, iY : Y })\n \n # Cada 25 iteraciones, imprimimos métricas.\n if step % 25 == 0: \n \n # Cálculo del accuracy.\n acc = np.mean(np.round(_pY) == Y)\n \n # Impresión de métricas.\n print('Step', step, '/', n_steps, '- Loss = ', _loss, '- Acc =', acc)\n \n # Obtenemos predicciones para cada punto de nuestro mapa de predicción _pX.\n _pY = sess.run(pY, feed_dict={ iX : _pX }).reshape((res, res))\n\n # Y lo guardamos para visualizar la animación.\n iPY.append(_pY)\n \n \n# ----- CÓDIGO ANIMACIÓN ----- #\n\nims = []\n\nfig = plt.figure(figsize=(10, 10))\n\nprint(\"--- Generando animación ---\")\n\nfor fr in range(len(iPY)):\n \n im = plt.pcolormesh(_x0, _x1, iPY[fr], cmap=\"coolwarm\", animated=True)\n\n # Visualización de la nube de datos.\n plt.scatter(X[Y == 0,0], X[Y == 0,1], c=\"skyblue\")\n plt.scatter(X[Y == 1,0], X[Y == 1,1], c=\"salmon\")\n\n # plt.title(\"Resultado Clasificación\")\n plt.tick_params(labelbottom=False, labelleft=False)\n\n ims.append([im])\n\nani = animation.ArtistAnimation(fig, ims, interval=50, blit=True, repeat_delay=1000)\n\nHTML(ani.to_html5_video())", "Step 0 / 1000 - Loss = 0.29063216 - Acc = 0.562\nStep 25 / 1000 - Loss = 0.18204297 - Acc = 0.632\nStep 50 / 1000 - Loss = 0.1471082 - Acc = 0.79\nStep 75 / 1000 - Loss = 0.13354021 - Acc = 0.854\nStep 100 / 1000 - Loss = 0.122594796 - Acc = 0.902\nStep 125 / 1000 - Loss = 0.111153014 - Acc = 0.942\nStep 150 / 1000 - Loss = 0.09965967 - Acc = 0.956\nStep 175 / 1000 - Loss = 0.08899061 - Acc = 0.968\nStep 200 / 1000 - Loss = 0.07819476 - Acc = 0.98\nStep 225 / 1000 - Loss = 0.06843367 - Acc = 0.984\nStep 250 / 1000 - Loss = 0.059809823 - Acc = 0.992\nStep 275 / 1000 - Loss = 0.051961992 - Acc = 0.994\nStep 300 / 1000 - Loss = 0.04506078 - Acc = 0.996\nStep 325 / 1000 - Loss = 0.03921565 - Acc = 0.998\nStep 350 / 1000 - Loss = 0.034442045 - Acc = 1.0\nStep 375 / 1000 - Loss = 0.030614918 - Acc = 1.0\nStep 400 / 1000 - Loss = 0.027491104 - Acc = 1.0\nStep 425 / 1000 - Loss = 0.024817096 - Acc = 1.0\nStep 450 / 1000 - Loss = 0.022489877 - Acc = 1.0\nStep 475 / 1000 - Loss = 0.020462362 - Acc = 1.0\nStep 500 / 1000 - Loss = 0.018674525 - Acc = 1.0\nStep 525 / 1000 - Loss = 0.01712375 - Acc = 1.0\nStep 550 / 1000 - Loss = 0.015800532 - Acc = 1.0\nStep 575 / 1000 - Loss = 0.014612312 - Acc = 1.0\nStep 600 / 1000 - Loss = 0.013566933 - Acc = 1.0\nStep 625 / 1000 - Loss = 0.012653999 - Acc = 1.0\nStep 650 / 1000 - Loss = 0.011832824 - Acc = 1.0\nStep 675 / 1000 - Loss = 0.011105223 - Acc = 1.0\nStep 700 / 1000 - Loss = 0.010456048 - Acc = 1.0\nStep 725 / 1000 - Loss = 0.009875296 - Acc = 1.0\nStep 750 / 1000 - Loss = 0.009351645 - Acc = 1.0\nStep 775 / 1000 - Loss = 0.008877279 - Acc = 1.0\nStep 800 / 1000 - Loss = 0.0084480485 - Acc = 1.0\nStep 825 / 1000 - Loss = 0.008039233 - Acc = 1.0\nStep 850 / 1000 - Loss = 0.0076591335 - Acc = 1.0\nStep 875 / 1000 - Loss = 0.0073076617 - Acc = 1.0\nStep 900 / 1000 - Loss = 0.0069831987 - Acc = 1.0\nStep 925 / 1000 - Loss = 0.0066823033 - Acc = 1.0\nStep 950 / 1000 - Loss = 0.0064092586 - Acc = 1.0\nStep 975 / 1000 - Loss = 0.0061591878 - Acc = 1.0\n--- Generando animación ---\n" ] ], [ [ "## Keras", "_____no_output_____" ] ], [ [ "import tensorflow as tf\nimport tensorflow.keras as kr\n\nfrom IPython.core.display import display, HTML\n\n\nlr = 0.01 # learning rate\nnn = [2, 16, 8, 1] # número de neuronas por capa.\n\n\n# Creamos el objeto que contendrá a nuestra red neuronal, como\n# secuencia de capas.\nmodel = kr.Sequential()\n\n# Añadimos la capa 1\nl1 = model.add(kr.layers.Dense(nn[1], activation='relu'))\n\n# Añadimos la capa 2\nl2 = model.add(kr.layers.Dense(nn[2], activation='relu'))\n\n# Añadimos la capa 3\nl3 = model.add(kr.layers.Dense(nn[3], activation='sigmoid'))\n\n# Compilamos el modelo, definiendo la función de coste y el optimizador.\nmodel.compile(loss='mse', optimizer=kr.optimizers.SGD(lr=0.05), metrics=['acc'])\n\n# Y entrenamos al modelo. Los callbacks \nmodel.fit(X, Y, epochs=100)", "Epoch 1/100\n500/500 [==============================] - 0s 111us/sample - loss: 0.2468 - acc: 0.5040\nEpoch 2/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.2457 - acc: 0.5100\nEpoch 3/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.2446 - acc: 0.5040\nEpoch 4/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.2434 - acc: 0.5160\nEpoch 5/100\n500/500 [==============================] - 0s 36us/sample - loss: 0.2422 - acc: 0.5200\nEpoch 6/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.2412 - acc: 0.5400\nEpoch 7/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.2400 - acc: 0.5460\nEpoch 8/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.2388 - acc: 0.5780\nEpoch 9/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.2376 - acc: 0.5840\nEpoch 10/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.2363 - acc: 0.5960\nEpoch 11/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.2350 - acc: 0.6280\nEpoch 12/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.2336 - acc: 0.6220\nEpoch 13/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.2320 - acc: 0.6280\nEpoch 14/100\n500/500 [==============================] - 0s 36us/sample - loss: 0.2305 - acc: 0.6580\nEpoch 15/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.2289 - acc: 0.6640\nEpoch 16/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.2272 - acc: 0.7040\nEpoch 17/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.2255 - acc: 0.7140\nEpoch 18/100\n500/500 [==============================] - 0s 47us/sample - loss: 0.2238 - acc: 0.7280\nEpoch 19/100\n500/500 [==============================] - 0s 42us/sample - loss: 0.2221 - acc: 0.7440\nEpoch 20/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.2201 - acc: 0.7620\nEpoch 21/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.2181 - acc: 0.7740\nEpoch 22/100\n500/500 [==============================] - 0s 50us/sample - loss: 0.2161 - acc: 0.7900\nEpoch 23/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.2140 - acc: 0.8040\nEpoch 24/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.2119 - acc: 0.8020\nEpoch 25/100\n500/500 [==============================] - 0s 48us/sample - loss: 0.2095 - acc: 0.8440\nEpoch 26/100\n500/500 [==============================] - 0s 44us/sample - loss: 0.2072 - acc: 0.8280\nEpoch 27/100\n500/500 [==============================] - 0s 43us/sample - loss: 0.2048 - acc: 0.8620\nEpoch 28/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.2023 - acc: 0.8720\nEpoch 29/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.1997 - acc: 0.8700\nEpoch 30/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.1970 - acc: 0.8940\nEpoch 31/100\n500/500 [==============================] - 0s 43us/sample - loss: 0.1941 - acc: 0.9100\nEpoch 32/100\n500/500 [==============================] - 0s 46us/sample - loss: 0.1910 - acc: 0.9220\nEpoch 33/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.1877 - acc: 0.9260\nEpoch 34/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.1843 - acc: 0.9380\nEpoch 35/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.1811 - acc: 0.9360\nEpoch 36/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.1779 - acc: 0.9480\nEpoch 37/100\n500/500 [==============================] - 0s 44us/sample - loss: 0.1748 - acc: 0.9520\nEpoch 38/100\n500/500 [==============================] - 0s 45us/sample - loss: 0.1713 - acc: 0.9580\nEpoch 39/100\n500/500 [==============================] - 0s 47us/sample - loss: 0.1680 - acc: 0.9620\nEpoch 40/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.1646 - acc: 0.9660\nEpoch 41/100\n500/500 [==============================] - 0s 44us/sample - loss: 0.1610 - acc: 0.9660\nEpoch 42/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.1576 - acc: 0.9720\nEpoch 43/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.1538 - acc: 0.9760\nEpoch 44/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.1503 - acc: 0.9740\nEpoch 45/100\n500/500 [==============================] - 0s 42us/sample - loss: 0.1463 - acc: 0.9780\nEpoch 46/100\n500/500 [==============================] - 0s 44us/sample - loss: 0.1426 - acc: 0.9800\nEpoch 47/100\n500/500 [==============================] - 0s 44us/sample - loss: 0.1386 - acc: 0.9840\nEpoch 48/100\n500/500 [==============================] - 0s 46us/sample - loss: 0.1346 - acc: 0.9860\nEpoch 49/100\n500/500 [==============================] - 0s 48us/sample - loss: 0.1305 - acc: 0.9920\nEpoch 50/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.1263 - acc: 0.9960\nEpoch 51/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.1221 - acc: 0.9980\nEpoch 52/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.1180 - acc: 1.0000\nEpoch 53/100\n500/500 [==============================] - 0s 36us/sample - loss: 0.1137 - acc: 0.9980\nEpoch 54/100\n500/500 [==============================] - 0s 42us/sample - loss: 0.1094 - acc: 1.0000\nEpoch 55/100\n500/500 [==============================] - 0s 43us/sample - loss: 0.1051 - acc: 1.0000\nEpoch 56/100\n500/500 [==============================] - 0s 44us/sample - loss: 0.1011 - acc: 1.0000\nEpoch 57/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.0969 - acc: 1.0000\nEpoch 58/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0930 - acc: 1.0000\nEpoch 59/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.0891 - acc: 1.0000\nEpoch 60/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0855 - acc: 1.0000\nEpoch 61/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.0819 - acc: 1.0000\nEpoch 62/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.0785 - acc: 1.0000\nEpoch 63/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0752 - acc: 1.0000\nEpoch 64/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.0719 - acc: 1.0000\nEpoch 65/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0689 - acc: 1.0000\nEpoch 66/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.0660 - acc: 1.0000\nEpoch 67/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0632 - acc: 1.0000\nEpoch 68/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.0606 - acc: 1.0000\nEpoch 69/100\n500/500 [==============================] - 0s 44us/sample - loss: 0.0581 - acc: 1.0000\nEpoch 70/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.0556 - acc: 1.0000\nEpoch 71/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.0533 - acc: 1.0000\nEpoch 72/100\n500/500 [==============================] - 0s 47us/sample - loss: 0.0511 - acc: 1.0000\nEpoch 73/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.0491 - acc: 1.0000\nEpoch 74/100\n500/500 [==============================] - 0s 43us/sample - loss: 0.0471 - acc: 1.0000\nEpoch 75/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.0452 - acc: 1.0000\nEpoch 76/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.0434 - acc: 1.0000\nEpoch 77/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.0418 - acc: 1.0000\nEpoch 78/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.0401 - acc: 1.0000\nEpoch 79/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.0386 - acc: 1.0000\nEpoch 80/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.0371 - acc: 1.0000\nEpoch 81/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.0358 - acc: 1.0000\nEpoch 82/100\n500/500 [==============================] - 0s 51us/sample - loss: 0.0344 - acc: 1.0000\nEpoch 83/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.0332 - acc: 1.0000\nEpoch 84/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.0320 - acc: 1.0000\nEpoch 85/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.0309 - acc: 1.0000\nEpoch 86/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.0298 - acc: 1.0000\nEpoch 87/100\n500/500 [==============================] - 0s 41us/sample - loss: 0.0288 - acc: 1.0000\nEpoch 88/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0278 - acc: 1.0000\nEpoch 89/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0269 - acc: 1.0000\nEpoch 90/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0260 - acc: 1.0000\nEpoch 91/100\n500/500 [==============================] - 0s 42us/sample - loss: 0.0251 - acc: 1.0000\nEpoch 92/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.0243 - acc: 1.0000\nEpoch 93/100\n500/500 [==============================] - 0s 42us/sample - loss: 0.0235 - acc: 1.0000\nEpoch 94/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0228 - acc: 1.0000\nEpoch 95/100\n500/500 [==============================] - 0s 43us/sample - loss: 0.0221 - acc: 1.0000\nEpoch 96/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0215 - acc: 1.0000\nEpoch 97/100\n500/500 [==============================] - 0s 39us/sample - loss: 0.0208 - acc: 1.0000\nEpoch 98/100\n500/500 [==============================] - 0s 38us/sample - loss: 0.0202 - acc: 1.0000\nEpoch 99/100\n500/500 [==============================] - 0s 40us/sample - loss: 0.0196 - acc: 1.0000\nEpoch 100/100\n500/500 [==============================] - 0s 37us/sample - loss: 0.0190 - acc: 1.0000\n" ] ], [ [ "## Sklearn", "_____no_output_____" ] ], [ [ "import sklearn as sk\nimport sklearn.neural_network\n\nfrom IPython.core.display import display, HTML\n\n\nlr = 0.01 # learning rate\nnn = [2, 16, 8, 1] # número de neuronas por capa.\n\n# Creamos el objeto del modelo de red neuronal multicapa.\nclf = sk.neural_network.MLPRegressor(solver='sgd', \n learning_rate_init=lr, \n hidden_layer_sizes=tuple(nn[1:]),\n verbose=True,\n n_iter_no_change=1000,\n batch_size = 64)\n\n\n# Y lo entrenamos con nuestro datos.\nclf.fit(X, Y)", "Iteration 1, loss = 0.66391606\nIteration 2, loss = 0.29448667\nIteration 3, loss = 0.13429471\nIteration 4, loss = 0.13165037\nIteration 5, loss = 0.13430276\nIteration 6, loss = 0.12556423\nIteration 7, loss = 0.12292571\nIteration 8, loss = 0.12204933\nIteration 9, loss = 0.12175702\nIteration 10, loss = 0.12129750\nIteration 11, loss = 0.12073281\nIteration 12, loss = 0.12028767\nIteration 13, loss = 0.11983928\nIteration 14, loss = 0.11939207\nIteration 15, loss = 0.11909108\nIteration 16, loss = 0.11836549\nIteration 17, loss = 0.11771654\nIteration 18, loss = 0.11703195\nIteration 19, loss = 0.11636100\nIteration 20, loss = 0.11559426\nIteration 21, loss = 0.11475135\nIteration 22, loss = 0.11391514\nIteration 23, loss = 0.11296898\nIteration 24, loss = 0.11183055\nIteration 25, loss = 0.11070522\nIteration 26, loss = 0.10945900\nIteration 27, loss = 0.10807801\nIteration 28, loss = 0.10653328\nIteration 29, loss = 0.10483565\nIteration 30, loss = 0.10299502\nIteration 31, loss = 0.10109678\nIteration 32, loss = 0.09868998\nIteration 33, loss = 0.09625805\nIteration 34, loss = 0.09356631\nIteration 35, loss = 0.09051455\nIteration 36, loss = 0.08720220\nIteration 37, loss = 0.08356481\nIteration 38, loss = 0.07956683\nIteration 39, loss = 0.07531904\nIteration 40, loss = 0.07057397\nIteration 41, loss = 0.06558903\nIteration 42, loss = 0.06042358\nIteration 43, loss = 0.05469307\nIteration 44, loss = 0.04914731\nIteration 45, loss = 0.04335916\nIteration 46, loss = 0.03773750\nIteration 47, loss = 0.03259090\nIteration 48, loss = 0.02785086\nIteration 49, loss = 0.02361280\nIteration 50, loss = 0.02024953\nIteration 51, loss = 0.01725262\nIteration 52, loss = 0.01477666\nIteration 53, loss = 0.01294427\nIteration 54, loss = 0.01150503\nIteration 55, loss = 0.01036948\nIteration 56, loss = 0.00950101\nIteration 57, loss = 0.00882902\nIteration 58, loss = 0.00840145\nIteration 59, loss = 0.00797252\nIteration 60, loss = 0.00769409\nIteration 61, loss = 0.00743825\nIteration 62, loss = 0.00731249\nIteration 63, loss = 0.00711639\nIteration 64, loss = 0.00700836\nIteration 65, loss = 0.00685081\nIteration 66, loss = 0.00670581\nIteration 67, loss = 0.00659854\nIteration 68, loss = 0.00655539\nIteration 69, loss = 0.00642937\nIteration 70, loss = 0.00637203\nIteration 71, loss = 0.00619710\nIteration 72, loss = 0.00614971\nIteration 73, loss = 0.00593245\nIteration 74, loss = 0.00579465\nIteration 75, loss = 0.00565489\nIteration 76, loss = 0.00553982\nIteration 77, loss = 0.00541618\nIteration 78, loss = 0.00532437\nIteration 79, loss = 0.00525496\nIteration 80, loss = 0.00514261\nIteration 81, loss = 0.00511693\nIteration 82, loss = 0.00499175\nIteration 83, loss = 0.00497192\nIteration 84, loss = 0.00491734\nIteration 85, loss = 0.00470830\nIteration 86, loss = 0.00461381\nIteration 87, loss = 0.00455140\nIteration 88, loss = 0.00446001\nIteration 89, loss = 0.00440248\nIteration 90, loss = 0.00430629\nIteration 91, loss = 0.00427582\nIteration 92, loss = 0.00420453\nIteration 93, loss = 0.00413087\nIteration 94, loss = 0.00406708\nIteration 95, loss = 0.00399991\nIteration 96, loss = 0.00394088\nIteration 97, loss = 0.00390739\nIteration 98, loss = 0.00384822\nIteration 99, loss = 0.00379567\nIteration 100, loss = 0.00372736\nIteration 101, loss = 0.00364839\nIteration 102, loss = 0.00359586\nIteration 103, loss = 0.00356903\nIteration 104, loss = 0.00350804\nIteration 105, loss = 0.00346888\nIteration 106, loss = 0.00341325\nIteration 107, loss = 0.00338402\nIteration 108, loss = 0.00334556\nIteration 109, loss = 0.00331617\nIteration 110, loss = 0.00327267\nIteration 111, loss = 0.00322546\nIteration 112, loss = 0.00316221\nIteration 113, loss = 0.00311790\nIteration 114, loss = 0.00308636\nIteration 115, loss = 0.00305983\nIteration 116, loss = 0.00307628\nIteration 117, loss = 0.00302102\nIteration 118, loss = 0.00299013\nIteration 119, loss = 0.00294987\nIteration 120, loss = 0.00295874\nIteration 121, loss = 0.00292606\nIteration 122, loss = 0.00289585\nIteration 123, loss = 0.00288184\nIteration 124, loss = 0.00286175\nIteration 125, loss = 0.00284965\nIteration 126, loss = 0.00286328\nIteration 127, loss = 0.00283168\nIteration 128, loss = 0.00285682\nIteration 129, loss = 0.00279665\nIteration 130, loss = 0.00278923\nIteration 131, loss = 0.00278239\nIteration 132, loss = 0.00276704\nIteration 133, loss = 0.00275697\nIteration 134, loss = 0.00275890\nIteration 135, loss = 0.00275535\nIteration 136, loss = 0.00282983\nIteration 137, loss = 0.00275359\nIteration 138, loss = 0.00272988\nIteration 139, loss = 0.00269894\nIteration 140, loss = 0.00272954\nIteration 141, loss = 0.00268760\nIteration 142, loss = 0.00267833\nIteration 143, loss = 0.00267846\nIteration 144, loss = 0.00269751\nIteration 145, loss = 0.00266955\nIteration 146, loss = 0.00265685\nIteration 147, loss = 0.00268063\nIteration 148, loss = 0.00265680\nIteration 149, loss = 0.00263361\nIteration 150, loss = 0.00262043\nIteration 151, loss = 0.00262108\nIteration 152, loss = 0.00262173\nIteration 153, loss = 0.00263316\nIteration 154, loss = 0.00259775\nIteration 155, loss = 0.00258960\nIteration 156, loss = 0.00263879\nIteration 157, loss = 0.00259500\nIteration 158, loss = 0.00257932\nIteration 159, loss = 0.00259434\nIteration 160, loss = 0.00256704\nIteration 161, loss = 0.00258173\nIteration 162, loss = 0.00253499\nIteration 163, loss = 0.00253539\nIteration 164, loss = 0.00253766\nIteration 165, loss = 0.00255039\nIteration 166, loss = 0.00253523\nIteration 167, loss = 0.00253166\nIteration 168, loss = 0.00252858\nIteration 169, loss = 0.00253196\nIteration 170, loss = 0.00251232\nIteration 171, loss = 0.00252011\nIteration 172, loss = 0.00251934\nIteration 173, loss = 0.00249041\nIteration 174, loss = 0.00249983\nIteration 175, loss = 0.00249816\nIteration 176, loss = 0.00249634\nIteration 177, loss = 0.00249739\nIteration 178, loss = 0.00249030\nIteration 179, loss = 0.00246445\nIteration 180, loss = 0.00250390\nIteration 181, loss = 0.00247568\nIteration 182, loss = 0.00247083\nIteration 183, loss = 0.00247611\nIteration 184, loss = 0.00246227\nIteration 185, loss = 0.00245628\nIteration 186, loss = 0.00245701\nIteration 187, loss = 0.00246615\nIteration 188, loss = 0.00244919\nIteration 189, loss = 0.00245754\nIteration 190, loss = 0.00245784\nIteration 191, loss = 0.00243623\nIteration 192, loss = 0.00245733\nIteration 193, loss = 0.00245661\nIteration 194, loss = 0.00242044\nIteration 195, loss = 0.00241922\nIteration 196, loss = 0.00242431\nIteration 197, loss = 0.00242330\nIteration 198, loss = 0.00245886\nIteration 199, loss = 0.00242789\nIteration 200, loss = 0.00240292\n" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import plotly.graph_objs as go\nimport numpy as np", "_____no_output_____" ], [ "fig = go.FigureWidget(data=[])", "_____no_output_____" ], [ "fig.add_scatter(\n x=[],\n y=[],\n mode='lines'\n)", "_____no_output_____" ], [ "fig.add_scatter(\n x=[],\n y=[],\n mode='lines'\n)", "_____no_output_____" ], [ "with fig.batch_update():\n fig.data[0].x = np.arange(100)\n fig.data[0].y = np.arange(100)*2", "_____no_output_____" ], [ "with fig.batch_update():\n fig.data[1].x = np.arange(100)\n fig.data[1].y = np.arange(100)*3", "_____no_output_____" ], [ "fig.data", "_____no_output_____" ], [ "foo=np.empty([0,0])", "_____no_output_____" ], [ "len(fig.data)", "_____no_output_____" ], [ "x = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]\nx_rev = x[::-1]\n\ny1_upper = [2, 3, 4, 5, 6, 7, 8, 9, 10, 11]\ny1_lower = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]\ny1_lower = y1_lower[::-1]\ny1_upper + y1_lower", "_____no_output_____" ], [ "x+x_rev", "_____no_output_____" ], [ "np.hstack((np.array(y1_upper),np.array(y1_lower)))", "_____no_output_____" ], [ "fig=go.FigureWidget(data=[])\nfig.add_scatter(\n x=[],\n y=[],\n mode='lines',\n showlegend=False,\n fill='tozerox',\n line=dict(color='rgba(255,255,255,0)'),\n fillcolor='rgba(255,0,0,0.2)'\n)\nfig.add_scatter(\n x=[],\n y=[],\n mode='lines',\n)\nwith fig.batch_update():\n fig.data[0].x = x+x_rev\n fig.data[0].y = y1_upper + y1_lower\n y_1 = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]\nwith fig.batch_update():\n fig.data[1].x = x\n fig.data[1].y = y_1\n fig.show()", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Arctic Project in Linear Regression: K-fold + Y:Area", "_____no_output_____" ], [ "## Import libraries", "_____no_output_____" ] ], [ [ "library(MASS)\nlibrary(tidyverse)", "── \u001b[1mAttaching packages\u001b[22m ──────────────────────────────────────────────────── tidyverse 1.3.0 ──\n\n\u001b[32m✔\u001b[39m \u001b[34mggplot2\u001b[39m 3.3.2 \u001b[32m✔\u001b[39m \u001b[34mpurrr \u001b[39m 0.3.4\n\u001b[32m✔\u001b[39m \u001b[34mtibble \u001b[39m 3.0.4 \u001b[32m✔\u001b[39m \u001b[34mdplyr \u001b[39m 1.0.2\n\u001b[32m✔\u001b[39m \u001b[34mtidyr \u001b[39m 1.1.2 \u001b[32m✔\u001b[39m \u001b[34mstringr\u001b[39m 1.4.0\n\u001b[32m✔\u001b[39m \u001b[34mreadr \u001b[39m 1.4.0 \u001b[32m✔\u001b[39m \u001b[34mforcats\u001b[39m 0.5.0\n\n── \u001b[1mConflicts\u001b[22m ─────────────────────────────────────────────────────── tidyverse_conflicts() ──\n\u001b[31m✖\u001b[39m \u001b[34mdplyr\u001b[39m::\u001b[32mfilter()\u001b[39m masks \u001b[34mstats\u001b[39m::filter()\n\u001b[31m✖\u001b[39m \u001b[34mdplyr\u001b[39m::\u001b[32mlag()\u001b[39m masks \u001b[34mstats\u001b[39m::lag()\n\u001b[31m✖\u001b[39m \u001b[34mdplyr\u001b[39m::\u001b[32mselect()\u001b[39m masks \u001b[34mMASS\u001b[39m::select()\n\n" ] ], [ [ "## Load data", "_____no_output_____" ] ], [ [ "arctic <- read.csv(\"arctic_data.csv\",stringsAsFactors = F) ", "_____no_output_____" ] ], [ [ "## Data segmentation", "_____no_output_____" ] ], [ [ "folds <- cut(seq(1,nrow(arctic)), breaks = 10, labels = FALSE) ", "_____no_output_____" ] ], [ [ "## Prediction", "_____no_output_____" ] ], [ [ "prediction <- as.data.frame(\n sapply(1:10, FUN = function(i) # loop 1:K\n{\n testID <- which(folds == i, arr.ind = TRUE)\n test <- arctic[testID, ]\n train <- arctic[-testID, ] # set K-fold\n \n # print(test) # if needed\n \n # linear regression\n model <- lm(area~rainfall+daylight+population+CO2+ozone+ocean_temp+land_temp,data=train)\n \n # print(summary(model)) # if needed\n \n # prediction output\n predict(model,test)\n}))", "_____no_output_____" ] ], [ [ "## Table gathering and merging", "_____no_output_____" ] ], [ [ "pred_gather <- gather(data=prediction, key=\"fold\",value=\"prediction\",1:10)\nresult <- as.data.frame(cbind(arctic[,c(1,6)],pred_gather))", "_____no_output_____" ] ], [ [ "## Calculate value of R^2", "_____no_output_____" ] ], [ [ "result[\"R^2\"] <- ((result$area-result$prediction)^2)\nR_square <- sum(result$`R^2`)/490", "_____no_output_____" ] ], [ [ "## Plot line chart (Prediction vs True) with title, legend, and specific size of figure", "_____no_output_____" ] ], [ [ "{plot(result$observation,result$area,type ='l',ylim = c(0,1.5),lwd = '2',xlab = \"Date\", ylab = \"Value\",xaxt='n')\n lines(result$observation,result$prediction,lty=1,col='red',lwd = '2')\n axis(1,at=c(1,61,121,181,241,301,361,421,481),\n labels=c(\"Jan 1980\",\"Jan 1985\",\"Jan 1990\",\"Jan 1995\",\"Jan 2000\",\"Jan 2005\",\"Jan 2010\",\"Jan 2015\",\"Jan 2020\"))\n title(main = list(\"Linear Regression\", cex = 1.5, col = \"red\", font = 3))\n legend(\"topright\", inset=.05, c(\"Prediction\",\"True\"), bty = 'n', lty=c(1, 1), col=c(\"red\", \"black\"),lwd =c(2, 2))\n options(repr.plot.width=20, repr.plot.height=10)\n}", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "<a href=\"https://colab.research.google.com/github/aldo-arevalo/mimic-code/blob/master/notebooks/ICUglycemia/Notebooks/2_0_ara_pairing_II.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>", "_____no_output_____" ], [ "# Data extraction and Pairing of Insulin Inputs to Glucose Measurements in the ICU\n## Interactive notebook: Part II\n\nAuthors: [Aldo Robles Arévalo](mailto:[email protected]); Jason Maley; Lawrence Baker; Susana M. da Silva Vieira; João M. da Costa Sousa; Stan Finkelstein; Jesse D. Raffa; Roselyn Cristelle; Leo Celi; Francis DeMichele\n\n## Overview\n\nThis notebook contains the pairing of pre-processed glucose readings and insulin inputs from the Medical Information Mart for Intensive Care (MIMIC).The curation is detailed in *1.0-ara-data-curation-I.ipynb*.", "_____no_output_____" ], [ "## General instructions\nTo perform the queries, do not forget to specify your project ID that grants you access to the MIMIC database hosted in *bigQuery*. Substitute `projectid` variable with the name of that project. In case you want to save the dataframes to your *BigQuery* project, uncomment and substitute `your_dataset` with the name of your *BigQuery* dataset and execute.\n\nYou can also save the created dataframes and figures in your Google Drive account. After mounting your drive, substitute `base_dir` variable with the path of the folder where you want to save them. In this notebook that folder was named `Insulin Therapy ICU` and `MyDrive` is the parent folder. Figures are saved in the path *Insulin Therapy ICU/DataExtraction/MIMIC_III/Figures/*, you should change it according to your needs or create the folders with the exact names in your Google Drive.", "_____no_output_____" ], [ "## Pairing rules\nOnce merged the insulin inputs and glucose readings from the *1.0-ara-data-curation-I.ipynb* notebok, now we continue with the **pairing** of an insulin event with a preceding glucose reading.\n\nThe goal is to link each insulin dose with the nearest glucose measurement. For this complex task, the following rules were implemented. This operation is done in BigQuery. The following rules or assumptions are proposed:\n\n1. **Rule 1**: A glucose reading should precede a regular insulin administration by up to 90 minutes. This basis for this time window is derived from the diabetic ketoacidosis guidelines which recommend measuring glucose values every 60 minutes while receiving an insulin infusion. An additional 30 minutes were added, 90 minutes in total, to this interval to account for the time it may take for providers to register the event. \n2. **Rule 2**: When a regular insulin event is not preceded, but instead followed, by a blood glucose measurement, this glucose reading is paired with the regular insulin administration if they are recorded within 90 minutes of each other.\n3. **Rule 3**: Sometimes a regular insulin infusion/bolus appears between 2 blood glucose measurements. In this case, the higher glucose value is paired with the regular insulin entry as long as they are entered within 90 minutes of each other.\n4. **Rule 4**: When a regular insulin bolus occurs very close to a regular insulin infusion rate, it is assumed that the patient was given a bolus and then commenced on an infusion. Both regular insulin entries are paired with the preceding blood glucose measurement, or the posterior glucose reading in case its value is higher than the preceding blood glucose and is entered within 90 minutes of the insulin dose.\n5. No glucose values below 90 mg/dL is paired with a subsequent regular insulin bolus or infusion. No clinician will treat this low of a blood glucose value with a regular insulin bolus or infusion.", "_____no_output_____" ], [ "# Code\n\n## Import dependencies and libraries", "_____no_output_____" ] ], [ [ "import pandas as pd\nimport numpy as np\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nimport matplotlib.colors as colors\nfrom scipy import stats\nfrom datetime import datetime\nimport time\nimport warnings\n\n# Below imports are used to print out pretty pandas dataframes\nfrom IPython.display import display, HTML\n\n# Imports for accessing data using Google BigQuery.\nfrom google.cloud import bigquery\nfrom google.colab import files, auth\nauth.authenticate_user()\nprint('Authenticated')\n%load_ext google.colab.data_table\n\n# Function to submit query to BigQuery\ndef q(query,projectid):\n client = bigquery.Client(location=\"US\",project=projectid)\n # Location must match that of the dataset(s) referenced in the query.\n query_job = client.query(query,\n location=\"US\",)\n return query_job.to_dataframe()\n\n#Rounding (for heatmap categories)\ndef myround(x, base):\n return int(base * round(float(x)/base))\n\ndef convert_to_datetime(df,time_cols):\n for t_col in time_cols:\n df[t_col] = pd.to_datetime(df[t_col])\n \n return(df)", "Authenticated\n" ], [ "from google.colab import drive\ndrive.mount('/content/gdrive')\n# Select your own folder\nbase_dir = \"/content/gdrive/My Drive/Insulin Therapy ICU\"", "_____no_output_____" ] ], [ [ "## Adjusted datasets\n\n* **Note 1**: Substitute `your_dataset` with the name of your dataset ID (Line 850) where you hosted/stored the tables created in the `1.0-ara-pairing-I.ipynb` notebook. \n* **Note 2**: The table `glucose_insulin_ICU` was created in `1.0-ara-pairing-I.ipynb` notebook. It is equivalent to `glucose_insulin_ICU.csv`.", "_____no_output_____" ] ], [ [ "# Import dataset adjusted or aligned\nprojectid = \"YOUR_PROJECT_ID\" # <-- Add your project ID\n\nquery =\"\"\"\nWITH pg AS(\n SELECT p1.*\n\n -- Column GLC_AL that would gather paired glucose values according to the proposed rules\n ,(CASE\n -- 1ST CLAUSE\n -- When previous and following rows are glucose readings, select the glucose value that \n -- has the shortest time distance to insulin bolus/infusion.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Posterior glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n AND ( -- Preceding glucose\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.GLC,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 2ND CLAUSE\n -- In case the posterior glucose reading is higher than the preceding\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose measurements\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n -- Preceding glucose\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose\n AND ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose values is higher than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,1) OVER(w) \n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.GLC,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 3RD CLAUSE\n -- When previous timestamp is an insulin bolus/infusion event\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above and regular insulin\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LAG(p1.INSULINTYPE,2) OVER(w)) IN('Short')\n -- One row above there is another insulin event\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a shortime or equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Preceding glucose 2 rows above is equal or greater than 90 min\n AND (LAG(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the preceding glucose value 2 rows above\n THEN (LAG(p1.GLC,2) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 4TH CLAUSE\n -- When previous timestamp is for Insulin bolus/infusion but posterior glucose\n -- is higher than the preceding glucose 2 rows above.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row above there is another regular insulin\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,1) OVER(w), p1.timer, MINUTE)) <= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is higher than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) < LEAD(p1.GLC,1) OVER(w)\n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.GLC,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 5TH CLAUSE\n -- When posterior timestamp is for Insulin bolus/infusion but preceding is glucose\n -- and there is a glucose 2 rows below.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a shorter OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose reading is greater or equal to 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose value (2 rows below) is lower than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,2) OVER(w)\n -- Return the PRECEDING glucose (1 row above) measurement that gathers the previous conditions\n THEN (LAG(p1.GLC,1) OVER(w)) \n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 6TH CLAUSE\n -- When posterior glucose reading (2 rows below) is higher than preceding glucose.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another insulin event\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose (2 rows below) occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,2) OVER(w), p1.timer, MINUTE)) <= 90\n -- Preceding glucose 1 row above occures up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose value (2 rows below) is higher than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,2) OVER(w)\n -- Return the POSTERIOR glucose (2 rows below) measurement that gathers the previous conditions\n THEN (LEAD(p1.GLC,2) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 7TH CLAUSE\n -- When it is the last insulin dose and record in an ICU stay\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin, should be equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.GLC,1) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 8TH CLAUSE\n -- When there is no preceding glucose reading within 90 min, but there is a posterior \n -- glucose within 90 min\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin is greater than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > 90)\n -- Time-gap between posterior glucose and insulin is equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Return the POSTERIOR glucose (1 rows below) measurement that gathers the previous conditions\n THEN (LEAD(p1.GLC,1) OVER(w))\n\n \n -- Otherwise, return null value and finish CASE clause\n ELSE null END\n ) AS GLC_AL\n\n -- ---------------------------------------------------------------------------------------------\n -- Column GLCTIMER_AL that would gather the timestamp of the paired glucose reading\n , (CASE \n -- 1ST CLAUSE\n -- When previous and following rows are glucose readings,vselect the glucose value that \n -- has the shortest time distance to insulin bolus/infusion.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Posterior glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n AND ( -- Preceding glucose\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.TIMER,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 2ND CLAUSE\n -- In case the posterior glucose reading is higher than the preceding\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose measurements\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n -- Preceding glucose\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose\n AND ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose values is higher than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,1) OVER(w) \n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.TIMER,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 3RD CLAUSE\n -- When previous timestamp is an insulin bolus/infusion event\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above and regular insulin\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LAG(p1.INSULINTYPE,2) OVER(w)) IN('Short')\n -- One row above there is another insulin event\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a shortime or equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Preceding glucose 2 rows above is equal or greater than 90 min\n AND (LAG(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the preceding glucose value 2 rows above\n THEN (LAG(p1.TIMER,2) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 4TH CLAUSE\n -- When previous timestamp is for Insulin bolus/infusion but posterior glucose\n -- is higher than the preceding glucose 2 rows above.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row above there is another regular insulin\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,1) OVER(w), p1.timer, MINUTE)) <= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is higher than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) < LEAD(p1.GLC,1) OVER(w)\n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.TIMER,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 5TH CLAUSE\n -- When posterior timestamp is for Insulin bolus/infusion but preceding is glucose\n -- and there is a glucose 2 rows below.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a shorter OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose reading is greater or equal to 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose value (2 rows below) is lower than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,2) OVER(w)\n -- Return the PRECEDING glucose (1 row above) measurement that gathers the previous conditions\n THEN (LAG(p1.TIMER,1) OVER(w)) \n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 6TH CLAUSE\n -- When posterior glucose reading (2 rows below) is higher than preceding glucose.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose (2 rows below) occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,2) OVER(w), p1.timer, MINUTE)) <= 90\n -- Preceding glucose 1 row above occures up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose value (2 rows below) is higher than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,2) OVER(w)\n -- Return the POSTERIOR glucose (2 rows below) measurement that gathers the previous conditions\n THEN (LEAD(p1.TIMER,2) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 7TH CLAUSE\n -- When it is the last insulin dose and record in an ICU stay\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin, should be equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.TIMER,1) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 8TH CLAUSE\n -- When there is no preceding glucose reading within 90 min, but there is a posterior \n -- glucose within 90 min\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin is greater than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > 90)\n -- Time-gap between posterior glucose and insulin is equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Return the timestamp of the POSTERIOR glucose (1 rows below) measurement that gathers the \n -- previous conditions\n THEN (LEAD(p1.TIMER,1) OVER(w))\n\n -- Otherwise, return null value and finish CASE clause\n ELSE null END\n ) AS GLCTIMER_AL\n\n -- -----------------------------------------------------------------------------------------------\n -- Column GLCSOURCE_AL that would indicate whether is fingerstick or lab analyzer sample of \n -- the paired glucose reading\n , (CASE\n -- 1ST CLAUSE\n -- When previous and following rows are glucose readings,vselect the glucose value that \n -- has the shortest time distance to insulin bolus/infusion.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Posterior glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n AND ( -- Preceding glucose\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.GLCSOURCE,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 2ND CLAUSE\n -- In case the posterior glucose reading is higher than the preceding\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose measurements\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n -- Preceding glucose\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose\n AND ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose values is higher than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,1) OVER(w) \n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.GLCSOURCE,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 3RD CLAUSE\n -- When previous timestamp is an insulin bolus/infusion event\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above and regular insulin\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LAG(p1.INSULINTYPE,2) OVER(w)) IN('Short')\n -- One row above there is another insulin event\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a shortime or equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Preceding glucose 2 rows above is equal or greater than 90 min\n AND (LAG(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the preceding glucose value 2 rows above\n THEN (LAG(p1.GLCSOURCE,2) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 4TH CLAUSE\n -- When previous timestamp is for Insulin bolus/infusion but posterior glucose\n -- is higher than the preceding glucose 2 rows above.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row above there is another regular insulin\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,1) OVER(w), p1.timer, MINUTE)) <= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is higher than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) < LEAD(p1.GLC,1) OVER(w)\n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.GLCSOURCE,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 5TH CLAUSE\n -- When posterior timestamp is for Insulin bolus/infusion but preceding is glucose\n -- and there is a glucose 2 rows below.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a shorter OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose reading is greater or equal to 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose value (2 rows below) is lower than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,2) OVER(w)\n -- Return the PRECEDING glucose (1 row above) measurement that gathers the previous conditions\n THEN (LAG(p1.GLCSOURCE,1) OVER(w)) \n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 6TH CLAUSE\n -- When posterior glucose reading (2 rows below) is higher than preceding glucose.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose (2 rows below) occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,2) OVER(w), p1.timer, MINUTE)) <= 90\n -- Preceding glucose 1 row above occures up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose value (2 rows below) is higher than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,2) OVER(w)\n -- Return the POSTERIOR glucose (2 rows below) measurement that gathers the previous conditions\n THEN (LEAD(p1.GLCSOURCE,2) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 7TH CLAUSE\n -- When it is the last insulin dose and record in an ICU stay\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin, should be equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.GLCSOURCE,1) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 8TH CLAUSE\n -- When there is no preceding glucose reading within 90 min, but there is a posterior \n -- glucose within 90 min\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin is greater than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > 90)\n -- Time-gap between posterior glucose and insulin is equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Return the whether is figerstick or lab analyzer the POSTERIOR glucose (1 rows below) measurement \n -- that gathers the previous conditions\n THEN (LEAD(p1.GLCSOURCE,1) OVER(w))\n\n -- Otherwise, return null value and finish CASE clause\n ELSE null END\n ) AS GLCSOURCE_AL\n\n -- ---------------------------------------------------------------------------------------------\n -- Column RULE that indicateS which pairing rule is applied for the i^th case\n , (CASE\n -- 1ST CLAUSE\n -- When previous and following rows are glucose readings,vselect the glucose value that \n -- has the shortest time distance to insulin bolus/infusion.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Posterior glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n AND ( -- Preceding glucose\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN 1\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 2ND CLAUSE\n -- In case the posterior glucose reading is higher than the preceding\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding and posterior glucose measurements\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n -- Preceding glucose\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose\n AND ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose values is higher than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,1) OVER(w) \n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN 3\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 3RD CLAUSE\n -- When previous timestamp is an insulin bolus/infusion event\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above and regular insulin\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LAG(p1.INSULINTYPE,2) OVER(w)) IN('Short')\n -- One row above there is another insulin event\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a shortime or equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Preceding glucose 2 rows above is equal or greater than 90 min\n AND (LAG(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the preceding glucose value 2 rows above\n THEN 4\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 4TH CLAUSE\n -- When previous timestamp is for Insulin bolus/infusion but posterior glucose\n -- is higher than the preceding glucose 2 rows above.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row above there is another regular insulin\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,1) OVER(w), p1.timer, MINUTE)) <= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is higher than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) < LEAD(p1.GLC,1) OVER(w)\n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN 4\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 5TH CLAUSE\n -- When posterior timestamp is for Insulin bolus/infusion but preceding is glucose\n -- and there is a glucose 2 rows below.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a shorter OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose reading is greater or equal to 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose value (2 rows below) is lower than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,2) OVER(w)\n -- Return the PRECEDING glucose (1 row above) measurement that gathers the previous conditions\n THEN 4\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 6TH CLAUSE\n -- When posterior glucose reading (2 rows below) is higher than preceding glucose.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose (2 rows below) occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,2) OVER(w), p1.timer, MINUTE)) <= 90\n -- Preceding glucose 1 row above occures up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90\n -- Posterior glucose value (2 rows below) is higher than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,2) OVER(w)\n -- Return the POSTERIOR glucose (2 rows below) measurement that gathers the previous conditions\n THEN 4\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 7TH CLAUSE\n -- When it is the last insulin dose and record in an ICU stay\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin, should be equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN 1\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 8TH CLAUSE\n -- When there is no preceding glucose reading within 90 min, but there is a posterior \n -- glucose within 90 min\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin is greater than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > 90)\n -- Time-gap between posterior glucose and insulin is equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 90)\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Return the Rule number applied\n THEN 2\n \n -- Otherwise, return null value and finish CASE clause\n ELSE null END\n ) AS RULE\n\n FROM `your_dataset.glucose_insulin_ICU` AS p1\n WINDOW w AS(PARTITION BY CAST(p1.HADM_ID AS INT64) ORDER BY p1.TIMER)\n)\n\n-- Create a colum that identifies the glucose readings were paired and are duplicated in pg\nSELECT pg.*\n, (CASE\n WHEN pg.GLCSOURCE_AL IS null \n AND (LEAD(pg.GLCTIMER_AL,1) OVER(x) = pg.GLCTIMER)\n THEN 1 \n WHEN pg.GLCSOURCE_AL IS null \n AND (LAG(pg.GLCTIMER_AL,1) OVER(x) = pg.GLCTIMER)\n AND LAG(endtime,1) OVER(x) IS NOT null \n THEN 1\n ELSE null END) AS Repeated\nFROM pg\nWINDOW x AS(PARTITION BY ICUSTAY_ID ORDER BY pg.timer)\n\"\"\"\n\nICUinputs_adjusted = q(query,projectid)\n\ndel query\n\n# Convert dtypes\nICUinputs_adjusted[[\"Repeated\",\"INFXSTOP\",\"RULE\"]] = ICUinputs_adjusted[\n [\"Repeated\",\"INFXSTOP\",\"RULE\"]].apply(pd.to_numeric, errors='coerce')\n\n# Remove values that are repeated due to the SQL query\nICUinputs_adjusted = ICUinputs_adjusted[ICUinputs_adjusted['Repeated']!=1]\n\n# Get statistics\ndisplay(HTML('<h5>Contains the following information</h5>'))\nprint(\"Entries: {}\".format(ICUinputs_adjusted.shape[0]))\nprint(\"Patients: {}\".format(ICUinputs_adjusted['SUBJECT_ID'].nunique()))\nprint(\"Hospital admissions: {}\".format(ICUinputs_adjusted['HADM_ID'].nunique()))\nprint('ICU stays: {}'.format(ICUinputs_adjusted['ICUSTAY_ID'].nunique()))\n\n# Rules\ndisplay(HTML('<h5>Frequency of the rules</h5>'))\nprint(ICUinputs_adjusted['RULE'].value_counts())", "_____no_output_____" ] ], [ [ "### Boluses of short-acting insulin", "_____no_output_____" ] ], [ [ "# Filtering for only short insulin boluses and all sources of glucose\nshort_BOL_adjusted = ICUinputs_adjusted[\n (ICUinputs_adjusted['INSULINTYPE']==\"Short\") & \n (ICUinputs_adjusted['EVENT'].str.contains('BOLUS'))].copy()\n\n# Get statistics\ndisplay(HTML('<h5>Contains the following information</h5>'))\nprint(\"Entries: {}\".format(short_BOL_adjusted.shape[0]))\nprint(\"Patients: {}\".format(short_BOL_adjusted['SUBJECT_ID'].nunique()))\nprint(\"Hospital admissions: {}\".format(short_BOL_adjusted['HADM_ID'].nunique()))\nprint('ICU stays: {}'.format(short_BOL_adjusted['ICUSTAY_ID'].nunique()))\n\ndisplay(short_BOL_adjusted[['INPUT','GLC_AL']].describe())\n\n# Save as CSV file, uncomment and modify as needed.\n# short_BOL_adjusted.to_csv(base_dir+\"/DataExtraction/BolusesCUR.csv\", index=False, \n# encoding='utf8', header = True)", "_____no_output_____" ], [ "# Aligned and not aligned entries\ndisplay(HTML('<h2>Boluses entries of short-acting insulin<h2>'))\nprint(\"Entries that were aligned: {}\".format(\n short_BOL_adjusted.shape[0]-short_BOL_adjusted.loc[np.isnan(\n short_BOL_adjusted.RULE),'RULE'].shape[0]))\nprint(\"Entries that weren't aligned: {}\".format(\n short_BOL_adjusted.loc[np.isnan(short_BOL_adjusted.RULE),'RULE'].shape[0]))\nprint(\"Non-paired percentage: {:0.2f}%\".format(\n short_BOL_adjusted.loc[np.isnan(\n short_BOL_adjusted.RULE),'RULE'].shape[0]/short_BOL_adjusted.shape[0]*100))", "_____no_output_____" ], [ "warnings.simplefilter('ignore')\n\n# From Part1 Notebook\nP99_bol_s = 18.0\n\n# Heatmap\nshort_BOL_heat = short_BOL_adjusted.dropna(subset=['GLC_AL']).copy()\nshort_BOL_heat['A'] = ((short_BOL_heat['GLCTIMER_AL'] - \n short_BOL_heat['STARTTIME'])/pd.Timedelta('1 minute'))*60\nshort_BOL_heat=short_BOL_heat.set_index('A')\n\n#Define the cell size on the heat map\nglc_base=25\nins_base=2\n\n#Define heatmap limits\nxlow=0\nxhigh=P99_bol_s\nylow=90\nyhigh=400\nxhigh-=ins_base\n\n#create categories for constructing the heatmap\nshort_BOL_heat['glc_cat']=(short_BOL_heat['GLC_AL'].apply(\n lambda x: myround(x, glc_base))/glc_base)\nshort_BOL_heat['ins_cat']=(short_BOL_heat['INPUT'].apply(\n lambda x: myround(x, ins_base))/ins_base)\n\n#create dataframe for the heatmap using pivot_table\nheat_df=pd.pivot_table(short_BOL_heat, values='ICUSTAY_ID', index=['glc_cat']\n, columns=['ins_cat'], aggfunc='count')\n#trim the heatmap dataframe based on the lmits specificed\nheat_df=heat_df.loc[ylow/glc_base:yhigh/glc_base:,xlow/ins_base:xhigh/ins_base:]\n\n#create labels for the x and y ticks\nheat_xtick=np.arange(xlow, xhigh+ins_base*2, ins_base)\nheat_ytick=np.arange(ylow, yhigh+glc_base*1, glc_base)\n\n#plot heatmap\nsns.set(style=\"ticks\", font_scale=1.2)\nfig, ax = plt.subplots(1, 1, figsize = (12, 12))\nax=sns.heatmap(heat_df, robust=True, annot=True, cmap=\"BuPu\", fmt=\"2.0f\"\n , xticklabels=heat_xtick, yticklabels=heat_ytick\n , norm=colors.PowerNorm(gamma=1./2.))\n\n#titles\nplt.title(f\"Glucose readings prior to a bolus of short-acting insulin\\n(n={int(heat_df.sum().values.sum())})\", \n fontsize=25)\nplt.ylabel(\"Blood glucose (mg/dL)\", fontsize=20)\nplt.xlabel(\"Insulin dose (U)\", fontsize=20)\n\n#invert axis and offset labels\nax.invert_yaxis()\nax.set_yticks(np.arange(0, ((yhigh-ylow)/glc_base)+1))\nax.set_xticks(np.arange(0, ((xhigh-xlow)/ins_base)+2))\n\n# Save figure, uncomment if needed.\nfig.savefig(base_dir+'/DataExtraction/ShortBolusHeatMap.png', bbox_inches='tight',\n dpi=fig.dpi)", "_____no_output_____" ] ], [ [ "### Infusions of short-acting insulin", "_____no_output_____" ] ], [ [ "warnings.simplefilter('default')\n\n# Filtering for only short insulin infusions and all sources of glucose\nshort_INF_adjusted = ICUinputs_adjusted[\n (ICUinputs_adjusted['INSULINTYPE']==\"Short\") & \n (ICUinputs_adjusted['EVENT'].str.contains('INFUSION'))].copy()\n\n# Get statistics\ndisplay(HTML('<h5>Counts</h5>'))\nprint(\"Entries: {}\".format(short_INF_adjusted.shape[0]))\nprint(\"Patients: {}\".format(short_INF_adjusted['SUBJECT_ID'].nunique()))\nprint(\"Hospital admissions: {}\".format(short_INF_adjusted['HADM_ID'].nunique()))\nprint('ICU stays: {}'.format(short_INF_adjusted['ICUSTAY_ID'].nunique()))\n\ndisplay(short_INF_adjusted[['INPUT_HRS','GLC_AL']].describe())", "_____no_output_____" ], [ "warnings.simplefilter('ignore')\n\n# Heatmap\nshort_INF_heat = short_INF_adjusted.dropna(subset=['GLC_AL']).copy()\nshort_INF_heat['A'] = ((short_INF_heat['GLCTIMER_AL'] - \n short_INF_heat['STARTTIME'])/pd.Timedelta('1 minute'))*60\nshort_INF_heat=short_INF_heat.set_index('A')\n\n#Define the cell size on the heat map\nglc_base=25\nins_base=2\n\n#Define heatmap limits\nxlow=0\nxhigh=P99_bol_s\nylow=90\nyhigh=400\nxhigh-=ins_base\n\n#create categories for constructing the heatmap\nshort_INF_heat['glc_cat']=(short_INF_heat['GLC_AL'].apply(\n lambda x: myround(x, glc_base))/glc_base)\nshort_INF_heat['ins_cat']=(short_INF_heat['INPUT'].apply(\n lambda x: myround(x, ins_base))/ins_base)\n\n#create dataframe for the heatmap using pivot_table\nheat_df_i=pd.pivot_table(short_INF_heat, values='ICUSTAY_ID', index=['glc_cat']\n, columns=['ins_cat'], aggfunc='count')\n#trim the heatmap dataframe based on the lmits specificed\nheat_df_i=heat_df_i.loc[ylow/glc_base:yhigh/glc_base:,xlow/ins_base:xhigh/ins_base:]\n\n#create labels for the x and y ticks\nheat_xtick=np.arange(xlow, xhigh+ins_base*2, ins_base)\nheat_ytick=np.arange(ylow, yhigh+glc_base*1, glc_base)\n\n#plot heatmap\nsns.set(style=\"ticks\", font_scale=1.2)\nfig, ax = plt.subplots(1, 1, figsize = (12, 12))\nax=sns.heatmap(heat_df_i, robust=True, annot=True, cmap=\"BuPu\", fmt=\"2.0f\"\n , xticklabels=heat_xtick, yticklabels=heat_ytick\n , norm=colors.PowerNorm(gamma=1./2.))\n\n#titles\nplt.title(f\"Glucose readings prior to infusions of short-acting insulin\\n(n={int(heat_df_i.sum().values.sum())})\", \n fontsize=25)\nplt.ylabel(\"Blood glucose (mg/dL)\", fontsize=20)\nplt.xlabel(\"Insulin dose (U/hr)\", fontsize=20)\n\n#invert axis and offset labels\nax.invert_yaxis()\nax.set_yticks(np.arange(0, ((yhigh-ylow)/glc_base)+1))\nax.set_xticks(np.arange(0, ((xhigh-xlow)/ins_base)+2))\n\n# Save figure, uncomment if needed.\nfig.savefig(base_dir+'/DataExtraction/ShortInfxnHeatMap.png', \n bbox_inches='tight',dpi=fig.dpi)", "_____no_output_____" ] ], [ [ "### Boluses of intermediate-acting insulin", "_____no_output_____" ] ], [ [ "warnings.simplefilter('default')\n\n# Filtering for only short insulin infusions and all sources of glucose\ninter_BOL_adjusted = ICUinputs_adjusted[\n (ICUinputs_adjusted['INSULINTYPE']==\"Intermediate\") & \n (ICUinputs_adjusted['EVENT'].str.contains('BOLUS'))].copy()\n\n# Get statistics\ndisplay(HTML('<h5>Contains the following information</h5>'))\nprint(\"Entries: {}\".format(inter_BOL_adjusted.shape[0]))\nprint(\"Patients: {}\".format(inter_BOL_adjusted['SUBJECT_ID'].nunique()))\nprint(\"Hospital admissions: {}\".format(inter_BOL_adjusted['HADM_ID'].nunique()))\nprint('ICU stays: {}'.format(inter_BOL_adjusted['ICUSTAY_ID'].nunique()))\n\ndisplay(inter_BOL_adjusted[['INPUT','GLC_AL']].describe())", "_____no_output_____" ], [ "# Aligned and not aligned entries\ndisplay(HTML('<h2>Boluses entries of intermediate-acting insulin<h2>'))\nprint(\"Entries that were aligned: {}\".format(\n inter_BOL_adjusted.shape[0]-inter_BOL_adjusted.loc[np.isnan(\n inter_BOL_adjusted.RULE),'RULE'].shape[0]))\nprint(\"Entries that weren't aligned: {}\".format(\n inter_BOL_adjusted.loc[np.isnan(inter_BOL_adjusted.RULE),'RULE'].shape[0]))\nprint(\"Non-paired percentage: {:0.2f}%\".format(\n inter_BOL_adjusted.loc[np.isnan(\n inter_BOL_adjusted.RULE),'RULE'].shape[0]/inter_BOL_adjusted.shape[0]*100))", "_____no_output_____" ] ], [ [ "### Boluses of long-acting insulin", "_____no_output_____" ] ], [ [ "warnings.simplefilter('default')\n\n# Filtering for only short insulin infusions and all sources of glucose\nlong_BOL_adjusted = ICUinputs_adjusted[\n (ICUinputs_adjusted['INSULINTYPE']==\"Long\") & \n (ICUinputs_adjusted['EVENT'].str.contains('BOLUS'))].copy()\n\n# Get statistics\ndisplay(HTML('<h5>Contains the following information</h5>'))\nprint(\"Entries: {}\".format(long_BOL_adjusted.shape[0]))\nprint(\"Patients: {}\".format(long_BOL_adjusted['SUBJECT_ID'].nunique()))\nprint(\"Hospital admissions: {}\".format(long_BOL_adjusted['HADM_ID'].nunique()))\nprint('ICU stays: {}'.format(long_BOL_adjusted['ICUSTAY_ID'].nunique()))\n\ndisplay(long_BOL_adjusted[['INPUT','GLC_AL']].describe())", "_____no_output_____" ], [ "# Aligned and not aligned entries\ndisplay(HTML('<h2>Boluses entries of long-acting insulin<h2>'))\nprint(\"Entries that were aligned: {}\".format(\n long_BOL_adjusted.shape[0]-long_BOL_adjusted.loc[np.isnan(\n long_BOL_adjusted.RULE),'RULE'].shape[0]))\nprint(\"Entries that weren't aligned: {}\".format(\n long_BOL_adjusted.loc[np.isnan(long_BOL_adjusted.RULE),'RULE'].shape[0]))\nprint(\"Non-paired percentage: {:0.2f}%\".format(\n long_BOL_adjusted.loc[np.isnan(\n long_BOL_adjusted.RULE),'RULE'].shape[0]/long_BOL_adjusted.shape[0]*100))", "_____no_output_____" ] ], [ [ "## Non-adjusted datasets\nTo complement this analysis, and to show the difference between implementing and not implementing the proposed rules, three cohorts were created: a) no pairing rules applied, b) paired a glucose reading recorded within 60 minutes of the insulin event instead of 90 minutes, and c) pairing a glucose reading.\n", "_____no_output_____" ], [ "### Scenario C\nGlucose readings CURATED and insulin inputs CURATED but NO RULES\n\n* **Note 1**: Add the name of your dataset hosted in BigQuery (Line 45). \n* **Note 2**: The table `glucose_insulin_ICU` was created in `1.0-ara-pairing-I.ipynb` notebook. It is equivalent to `glucose_insulin_ICU.csv`.", "_____no_output_____" ] ], [ [ "# GLUCOSE READINGS CURATED AND INSULIN INPUTS CURATED but no RULES\n\nquery = \"\"\"\nSELECT pg.*\n , (CASE\n WHEN pg.GLCSOURCE_AL IS null \n AND (LEAD(pg.GLCTIMER_AL,1) OVER(PARTITION BY pg.ICUSTAY_ID ORDER BY pg.TIMER) = pg.GLCTIMER)\n THEN 1 \n WHEN pg.GLCSOURCE_AL IS null \n AND (LAG(pg.GLCTIMER_AL,1) OVER(PARTITION BY pg.ICUSTAY_ID ORDER BY pg.timer) = pg.GLCTIMER)\n AND LAG(endtime,1) OVER(PARTITION BY ICUSTAY_ID ORDER BY timer) IS NOT null \n THEN 1\n ELSE null END) AS Repeated\n FROM(SELECT p1.* \n , (CASE\n -- Select the previous glucose value regardless the time distance\n WHEN p1.EVENT IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK') \n THEN (LAG(p1.GLC,1) OVER(w))\n ELSE null END\n ) AS GLC_AL\n\n , (CASE\n -- Select the previous glucose value regardless the time distance\n WHEN p1.EVENT IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK') \n THEN (LAG(p1.TIMER,1) OVER(w))\n ELSE null END\n ) AS GLCTIMER_AL\n\n , (CASE\n -- Select the previous glucose value regardless the time distance\n WHEN p1.EVENT IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK') \n THEN (LAG(p1.GLCSOURCE,1) OVER(w))\n ELSE null END\n ) AS GLCSOURCE_AL\n , (CASE\n -- Select the previous glucose value regardless the time distance\n WHEN p1.EVENT IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK') \n THEN 1\n ELSE null END\n ) AS RULE\n FROM `your_dataset.glucose_insulin_ICU` AS p1\n WINDOW w AS(PARTITION BY CAST(p1.HADM_ID AS INT64) ORDER BY p1.TIMER)\n ) AS pg\n\"\"\"\n\nglc_curALins_cur = q(query,projectid)\n\nqwe = glc_curALins_cur[(glc_curALins_cur['INSULINTYPE']==\"Short\") & \n (glc_curALins_cur['EVENT'].str.contains('BOLUS'))].copy()\n\ndisplay(HTML('<h4>Statistics for both glucose readings and insulin inputs CURATED</h4>'))\nprint(\"Total entries: {}\".format(glc_curALins_cur.shape[0]))\ndisplay(qwe[['INPUT','GLC_AL']].describe())\n\ndisplay(HTML('<h5>Contains the following information (only for short-acting)</h5>'))\nprint(\"Boluses of short-acting insulin: {}\".format(qwe.shape[0]))\nprint(\"Patients: {} out of {}\".format(qwe['SUBJECT_ID'].nunique(),\n glc_curALins_cur['SUBJECT_ID'].nunique()))\nprint(\"Hospital admissions: {}\".format(qwe['HADM_ID'].nunique()))\nprint('ICU stays: {}'.format(qwe['ICUSTAY_ID'].nunique()))\n\n# Rules\ndisplay(HTML('<h5>Frequency of the rules</h5>'))\nprint(qwe['RULE'].value_counts())\n\n# Save as CSV file, uncomment and modify as needed.\n# qwe.to_csv(base_dir+\"/DataExtraction/BolusesCUR_nr.csv\", index=False,\n# encoding='utf8', header = True)\n\ndel query,qwe", "/usr/lib/python3.6/json/decoder.py:355: ResourceWarning: unclosed <ssl.SSLSocket fd=63, family=AddressFamily.AF_INET, type=2049, proto=6, laddr=('172.28.0.2', 52706), raddr=('74.125.142.95', 443)>\n obj, end = self.scan_once(s, idx)\n/usr/lib/python3.6/json/decoder.py:355: ResourceWarning: unclosed <ssl.SSLSocket fd=64, family=AddressFamily.AF_INET, type=2049, proto=6, laddr=('172.28.0.2', 52660), raddr=('74.125.20.95', 443)>\n obj, end = self.scan_once(s, idx)\n" ] ], [ [ "### Scenario B\nGlucose reading CURATED and inulin inputs CURATED paired with rules (60 min)\n\n* **Note 1**: Substitute `your_dataset` with the name of your dataset ID (Line 849) where you hosted/stored the tables created in the `1.0-ara-pairing-I.ipynb` notebook. \n* **Note 2**: The table `glucose_insulin_ICU` was created in `1.0-ara-pairing-I.ipynb` notebook. It is equivalent to `glucose_insulin_ICU.csv`.", "_____no_output_____" ] ], [ [ "# Import dataset adjusted or aligned with 60 min\n\nquery =\"\"\"\nWITH pg AS(\n SELECT p1.*\n\n -- Column GLC_AL that would gather paired glucose values according to the proposed rules\n ,(CASE\n -- 1ST CLAUSE\n -- When previous and following rows are glucose readings, select the glucose value that \n -- has the shortest time distance to insulin bolus/infusion.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Posterior glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n AND ( -- Preceding glucose\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.GLC,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 2ND CLAUSE\n -- In case the posterior glucose reading is higher than the preceding\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose measurements\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n -- Preceding glucose\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose\n AND ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose values is higher than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,1) OVER(w) \n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.GLC,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 3RD CLAUSE\n -- When previous timestamp is an insulin bolus/infusion event\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above and regular insulin\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LAG(p1.INSULINTYPE,2) OVER(w)) IN('Short')\n -- One row above there is another insulin event\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a shortime or equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Preceding glucose 2 rows above is equal or greater than 90 min\n AND (LAG(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the preceding glucose value 2 rows above\n THEN (LAG(p1.GLC,2) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 4TH CLAUSE\n -- When previous timestamp is for Insulin bolus/infusion but posterior glucose\n -- is higher than the preceding glucose 2 rows above.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row above there is another regular insulin\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,1) OVER(w), p1.timer, MINUTE)) <= 60\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is higher than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) < LEAD(p1.GLC,1) OVER(w)\n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.GLC,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 5TH CLAUSE\n -- When posterior timestamp is for Insulin bolus/infusion but preceding is glucose\n -- and there is a glucose 2 rows below.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a shorter OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose reading is greater or equal to 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose value (2 rows below) is lower than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,2) OVER(w)\n -- Return the PRECEDING glucose (1 row above) measurement that gathers the previous conditions\n THEN (LAG(p1.GLC,1) OVER(w)) \n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 6TH CLAUSE\n -- When posterior glucose reading (2 rows below) is higher than preceding glucose.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another insulin event\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose (2 rows below) occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,2) OVER(w), p1.timer, MINUTE)) <= 60\n -- Preceding glucose 1 row above occures up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose value (2 rows below) is higher than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,2) OVER(w)\n -- Return the POSTERIOR glucose (2 rows below) measurement that gathers the previous conditions\n THEN (LEAD(p1.GLC,2) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 7TH CLAUSE\n -- When it is the last insulin dose and record in an ICU stay\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin, should be equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.GLC,1) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 8TH CLAUSE\n -- When there is no preceding glucose reading within 90 min, but there is a posterior \n -- glucose within 90 min\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin is greater than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > 90)\n -- Time-gap between posterior glucose and insulin is equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Return the POSTERIOR glucose (1 rows below) measurement that gathers the previous conditions\n THEN (LEAD(p1.GLC,1) OVER(w))\n\n \n -- Otherwise, return null value and finish CASE clause\n ELSE null END\n ) AS GLC_AL\n\n -- ---------------------------------------------------------------------------------------------\n -- Column GLCTIMER_AL that would gather the timestamp of the paired glucose reading\n , (CASE \n -- 1ST CLAUSE\n -- When previous and following rows are glucose readings,vselect the glucose value that \n -- has the shortest time distance to insulin bolus/infusion.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Posterior glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n AND ( -- Preceding glucose\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.TIMER,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 2ND CLAUSE\n -- In case the posterior glucose reading is higher than the preceding\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose measurements\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n -- Preceding glucose\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose\n AND ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose values is higher than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,1) OVER(w) \n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.TIMER,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 3RD CLAUSE\n -- When previous timestamp is an insulin bolus/infusion event\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above and regular insulin\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LAG(p1.INSULINTYPE,2) OVER(w)) IN('Short')\n -- One row above there is another insulin event\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a shortime or equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Preceding glucose 2 rows above is equal or greater than 90 min\n AND (LAG(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the preceding glucose value 2 rows above\n THEN (LAG(p1.TIMER,2) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 4TH CLAUSE\n -- When previous timestamp is for Insulin bolus/infusion but posterior glucose\n -- is higher than the preceding glucose 2 rows above.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row above there is another regular insulin\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,1) OVER(w), p1.timer, MINUTE)) <= 60\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is higher than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) < LEAD(p1.GLC,1) OVER(w)\n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.TIMER,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 5TH CLAUSE\n -- When posterior timestamp is for Insulin bolus/infusion but preceding is glucose\n -- and there is a glucose 2 rows below.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a shorter OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose reading is greater or equal to 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose value (2 rows below) is lower than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,2) OVER(w)\n -- Return the PRECEDING glucose (1 row above) measurement that gathers the previous conditions\n THEN (LAG(p1.TIMER,1) OVER(w)) \n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 6TH CLAUSE\n -- When posterior glucose reading (2 rows below) is higher than preceding glucose.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose (2 rows below) occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,2) OVER(w), p1.timer, MINUTE)) <= 60\n -- Preceding glucose 1 row above occures up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose value (2 rows below) is higher than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,2) OVER(w)\n -- Return the POSTERIOR glucose (2 rows below) measurement that gathers the previous conditions\n THEN (LEAD(p1.TIMER,2) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 7TH CLAUSE\n -- When it is the last insulin dose and record in an ICU stay\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin, should be equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.TIMER,1) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 8TH CLAUSE\n -- When there is no preceding glucose reading within 90 min, but there is a posterior \n -- glucose within 90 min\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin is greater than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > 90)\n -- Time-gap between posterior glucose and insulin is equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Return the timestamp of the POSTERIOR glucose (1 rows below) measurement that gathers the \n -- previous conditions\n THEN (LEAD(p1.TIMER,1) OVER(w))\n\n -- Otherwise, return null value and finish CASE clause\n ELSE null END\n ) AS GLCTIMER_AL\n\n -- -----------------------------------------------------------------------------------------------\n -- Column GLCSOURCE_AL that would indicate whether is fingerstick or lab analyzer sample of \n -- the paired glucose reading\n , (CASE\n -- 1ST CLAUSE\n -- When previous and following rows are glucose readings,vselect the glucose value that \n -- has the shortest time distance to insulin bolus/infusion.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Posterior glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n AND ( -- Preceding glucose\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.GLCSOURCE,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 2ND CLAUSE\n -- In case the posterior glucose reading is higher than the preceding\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose measurements\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n -- Preceding glucose\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose\n AND ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose values is higher than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,1) OVER(w) \n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.GLCSOURCE,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 3RD CLAUSE\n -- When previous timestamp is an insulin bolus/infusion event\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above and regular insulin\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LAG(p1.INSULINTYPE,2) OVER(w)) IN('Short')\n -- One row above there is another insulin event\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a shortime or equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Preceding glucose 2 rows above is equal or greater than 90 min\n AND (LAG(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the preceding glucose value 2 rows above\n THEN (LAG(p1.GLCSOURCE,2) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 4TH CLAUSE\n -- When previous timestamp is for Insulin bolus/infusion but posterior glucose\n -- is higher than the preceding glucose 2 rows above.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row above there is another regular insulin\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,1) OVER(w), p1.timer, MINUTE)) <= 60\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is higher than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) < LEAD(p1.GLC,1) OVER(w)\n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN (LEAD(p1.GLCSOURCE,1) OVER(w))\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 5TH CLAUSE\n -- When posterior timestamp is for Insulin bolus/infusion but preceding is glucose\n -- and there is a glucose 2 rows below.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a shorter OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose reading is greater or equal to 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose value (2 rows below) is lower than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,2) OVER(w)\n -- Return the PRECEDING glucose (1 row above) measurement that gathers the previous conditions\n THEN (LAG(p1.GLCSOURCE,1) OVER(w)) \n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 6TH CLAUSE\n -- When posterior glucose reading (2 rows below) is higher than preceding glucose.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose (2 rows below) occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,2) OVER(w), p1.timer, MINUTE)) <= 60\n -- Preceding glucose 1 row above occures up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose value (2 rows below) is higher than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,2) OVER(w)\n -- Return the POSTERIOR glucose (2 rows below) measurement that gathers the previous conditions\n THEN (LEAD(p1.GLCSOURCE,2) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 7TH CLAUSE\n -- When it is the last insulin dose and record in an ICU stay\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin, should be equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN (LAG(p1.GLCSOURCE,1) OVER(w))\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 8TH CLAUSE\n -- When there is no preceding glucose reading within 90 min, but there is a posterior \n -- glucose within 90 min\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin is greater than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > 90)\n -- Time-gap between posterior glucose and insulin is equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Return the whether is figerstick or lab analyzer the POSTERIOR glucose (1 rows below) measurement \n -- that gathers the previous conditions\n THEN (LEAD(p1.GLCSOURCE,1) OVER(w))\n\n -- Otherwise, return null value and finish CASE clause\n ELSE null END\n ) AS GLCSOURCE_AL\n\n -- ---------------------------------------------------------------------------------------------\n -- Column RULE that indicateS which pairing rule is applied for the i^th case\n , (CASE\n -- 1ST CLAUSE\n -- When previous and following rows are glucose readings,vselect the glucose value that \n -- has the shortest time distance to insulin bolus/infusion.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding and posterior glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Posterior glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n AND ( -- Preceding glucose\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN 1\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 2ND CLAUSE\n -- In case the posterior glucose reading is higher than the preceding\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding and posterior glucose measurements\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Time-gap between glucose and insulin, should be equal or less than 90 minutes\n -- Preceding glucose\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose\n AND ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose values is higher than the preceding glucose\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,1) OVER(w) \n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN 3\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 3RD CLAUSE\n -- When previous timestamp is an insulin bolus/infusion event\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above and regular insulin\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND (LAG(p1.INSULINTYPE,2) OVER(w)) IN('Short')\n -- One row above there is another insulin event\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a shortime or equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Preceding glucose 2 rows above is equal or greater than 90 min\n AND (LAG(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose value is lower than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) >= LEAD(p1.GLC,1) OVER(w)\n -- Return the preceding glucose value 2 rows above\n THEN 4\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 4TH CLAUSE\n -- When previous timestamp is for Insulin bolus/infusion but posterior glucose\n -- is higher than the preceding glucose 2 rows above.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 2 rows above\n AND (LAG(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row above there is another regular insulin\n AND (LAG(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LAG(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,1) OVER(w), p1.timer, MINUTE)) <= 60\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Posterior glucose value is higher than the preceding glucose 2 rows above\n AND LAG(p1.GLC,2) OVER(w) < LEAD(p1.GLC,1) OVER(w)\n -- Return the POSTERIOR glucose measurement that gathers the previous conditions\n THEN 4\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 5TH CLAUSE\n -- When posterior timestamp is for Insulin bolus/infusion but preceding is glucose\n -- and there is a glucose 2 rows below.\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a shorter OR equal time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Preceding glucose reading is greater or equal to 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Preceding glucose 2 rows above occured up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose value (2 rows below) is lower than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) >= LEAD(p1.GLC,2) OVER(w)\n -- Return the PRECEDING glucose (1 row above) measurement that gathers the previous conditions\n THEN 4\n \n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 6TH CLAUSE\n -- When posterior glucose reading (2 rows below) is higher than preceding glucose.\n \n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 2 rows below\n AND (LEAD(p1.GLCSOURCE,2) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- One row BELOW there is another regular insulin\n AND (LEAD(p1.EVENT,1) OVER(w)) IN('BOLUS_INYECTION','BOLUS_PUSH','INFUSION')\n AND (LEAD(p1.INSULINTYPE,1) OVER(w)) IN('Short')\n AND ( -- Preceding glucose has a longer time-gap to insulin than the posterior\n ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > \n ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,2) OVER(w)), p1.TIMER, MINUTE))\n )\n -- Posterior glucose reading is greater or equal to 90 mg/dL\n AND (LEAD(p1.GLC,2) OVER(w)) >= 90\n -- Posterior glucose (2 rows below) occurs within 90 minutes\n AND ABS(TIMESTAMP_DIFF(LEAD(p1.timer,2) OVER(w), p1.timer, MINUTE)) <= 60\n -- Preceding glucose 1 row above occures up to 90 minutes before\n AND ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60\n -- Posterior glucose value (2 rows below) is higher than the preceding glucose 1 row above\n AND LAG(p1.GLC,1) OVER(w) < LEAD(p1.GLC,2) OVER(w)\n -- Return the POSTERIOR glucose (2 rows below) measurement that gathers the previous conditions\n THEN 4\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 7TH CLAUSE\n -- When it is the last insulin dose and record in an ICU stay\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Identify preceding glucose reading\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin, should be equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Preceding glucose should be equal or greater than 90 mg/dL\n AND (LAG(p1.GLC,1) OVER(w)) >= 90\n -- Return the PRECEDING glucose measurement that gathers the previous conditions\n THEN 1\n\n -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%\n -- 8TH CLAUSE\n -- When there is no preceding glucose reading within 90 min, but there is a posterior \n -- glucose within 90 min\n\n -- Identify an insulin event either bolus or infusion\n WHEN p1.EVENT IN('BOLUS_INYECTION', 'BOLUS_PUSH', 'INFUSION')\n -- Regular insulin or short-acting\n AND p1.INSULINTYPE IN('Short')\n -- Identify preceding glucose reading 1 row above\n AND (LAG(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Identify posterior glucose reading 1 row below\n AND (LEAD(p1.GLCSOURCE,1) OVER(w)) IN('BLOOD', 'FINGERSTICK')\n -- Time-gap between preceding glucose and insulin is greater than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LAG(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) > 90)\n -- Time-gap between posterior glucose and insulin is equal or less than 90 minutes\n AND (ABS(TIMESTAMP_DIFF((LEAD(p1.TIMER,1) OVER(w)), p1.TIMER, MINUTE)) <= 60)\n -- Posterior glucose should be equal or greater than 90 mg/dL\n AND (LEAD(p1.GLC,1) OVER(w)) >= 90\n -- Return the Rule number applied\n THEN 2\n \n -- Otherwise, return null value and finish CASE clause\n ELSE null END\n ) AS RULE\n\n FROM `your_dataset.glucose_insulin_ICU` AS p1\n WINDOW w AS(PARTITION BY CAST(p1.HADM_ID AS INT64) ORDER BY p1.TIMER)\n)\n\n-- Create a colum that identifies the glucose readings were paired and are duplicated in pg\nSELECT pg.*\n, (CASE\n WHEN pg.GLCSOURCE_AL IS null \n AND (LEAD(pg.GLCTIMER_AL,1) OVER(x) = pg.GLCTIMER)\n THEN 1 \n WHEN pg.GLCSOURCE_AL IS null \n AND (LAG(pg.GLCTIMER_AL,1) OVER(x) = pg.GLCTIMER)\n AND LAG(endtime,1) OVER(x) IS NOT null \n THEN 1\n ELSE null END) AS Repeated\nFROM pg\nWINDOW x AS(PARTITION BY ICUSTAY_ID ORDER BY pg.timer)\n\"\"\"\n\nICU60min_adjusted = q(query,projectid)\n\ndel query\n\n# Convert dtypes\nICU60min_adjusted[[\"Repeated\",\"INFXSTOP\",\"RULE\"]] = ICU60min_adjusted[\n [\"Repeated\",\"INFXSTOP\",\"RULE\"]].apply(pd.to_numeric, errors='coerce')\n\n# Remove values that are repeated due to the SQL query\nICU60min_adjusted = ICU60min_adjusted[ICU60min_adjusted['Repeated']!=1]\n\n# Get statistics\ndisplay(HTML('<h5>Contains the following information</h5>'))\nprint(\"Entries: {}\".format(ICU60min_adjusted.shape[0]))\nprint(\"Patients: {}\".format(ICU60min_adjusted['SUBJECT_ID'].nunique()))\nprint(\"Hospital admissions: {}\".format(ICU60min_adjusted['HADM_ID'].nunique()))\nprint('ICU stays: {}'.format(ICU60min_adjusted['ICUSTAY_ID'].nunique()))\n\n# Rules\ndisplay(HTML('<h5>Frequency of the rules</h5>'))\nprint(ICU60min_adjusted['RULE'].value_counts())", "/usr/lib/python3.6/json/decoder.py:355: ResourceWarning: unclosed <ssl.SSLSocket fd=80, family=AddressFamily.AF_INET, type=2049, proto=6, laddr=('172.28.0.2', 52770), raddr=('74.125.20.95', 443)>\n obj, end = self.scan_once(s, idx)\n/usr/lib/python3.6/json/decoder.py:355: ResourceWarning: unclosed <ssl.SSLSocket fd=79, family=AddressFamily.AF_INET, type=2049, proto=6, laddr=('172.28.0.2', 53808), raddr=('74.125.195.95', 443)>\n obj, end = self.scan_once(s, idx)\n" ] ], [ [ "#### Boluses of short-acting insulin", "_____no_output_____" ] ], [ [ "# Filtering for only short insulin boluses and all sources of glucose\nshort_BOL_60 = ICU60min_adjusted[(ICU60min_adjusted['INSULINTYPE']==\"Short\") & \n (ICU60min_adjusted['EVENT'].str.contains('BOLUS'))].copy()\n\n# Get statistics\ndisplay(HTML('<h5>Contains the following information</h5>'))\nprint(\"Entries: {}\".format(short_BOL_60.shape[0]))\nprint(\"Patients: {}\".format(short_BOL_60['SUBJECT_ID'].nunique()))\nprint(\"Hospital admissions: {}\".format(short_BOL_60['HADM_ID'].nunique()))\nprint('ICU stays: {}'.format(short_BOL_60['ICUSTAY_ID'].nunique()))\n\ndisplay(short_BOL_60[['INPUT','GLC_AL']].describe())\n\n# Save as CSV file, uncomment and modify as needed.\n# short_BOL_60.to_csv(base_dir+\"/DataExtraction/BolusesCUR_60.csv\", index=False,\n# encoding='utf8', header = True)", "_____no_output_____" ] ], [ [ "# End", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown", "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Loops and Conditions", "_____no_output_____" ], [ "loops provides the methods of iteration while condition allows or blocks the code execution when specified condition\nis meet.", "_____no_output_____" ], [ "For Loop and while Loop", "_____no_output_____" ] ], [ [ "L = ['apple', 'banana','kite','cellphone']\nfor item in L:\n print(item)", "apple\nbanana\nkite\ncellphone\n" ], [ "range(5), range(5,100), sum(range(100))", "_____no_output_____" ], [ "L=[]\nfor k in range(10):\n L.append(10*k)\nL", "_____no_output_____" ], [ "D = {}\nfor i in range(5):\n for j in range(5):\n if i == j :\n D.update({(i,j) : 10*i+j})\n elif i!=j :\n D.update({(i,j): 100*i+j})\nprint(D)", "{(0, 0): 0, (0, 1): 1, (0, 2): 2, (0, 3): 3, (0, 4): 4, (1, 0): 100, (1, 1): 11, (1, 2): 102, (1, 3): 103, (1, 4): 104, (2, 0): 200, (2, 1): 201, (2, 2): 22, (2, 3): 203, (2, 4): 204, (3, 0): 300, (3, 1): 301, (3, 2): 302, (3, 3): 33, (3, 4): 304, (4, 0): 400, (4, 1): 401, (4, 2): 402, (4, 3): 403, (4, 4): 44}\n" ], [ "for i, item in enumerate(['apple', 'banana','kite','cellphone']):\n print(\"The\",i,\"th element is:\", item)", "The 0 th element is: apple\nThe 1 th element is: banana\nThe 2 th element is: kite\nThe 3 th element is: cellphone\n" ], [ "A=[10*k**2+5*k+1 for k in range(10)]\nprint(A)", "[1, 16, 51, 106, 181, 276, 391, 526, 681, 856]\n" ], [ "AA=[[10*x**2+5*y+1 for x in range(3)] for y in range(3)]\nprint(AA)", "[[1, 11, 41], [6, 16, 46], [11, 21, 51]]\n" ], [ "for i in range(3):\n for j in range(3):\n print(\"The\", \"(\",i,\",\",j,\")\",\"th element is: \", AA[i][j])", "The ( 0 , 0 ) th element is: 1\nThe ( 0 , 1 ) th element is: 11\nThe ( 0 , 2 ) th element is: 41\nThe ( 1 , 0 ) th element is: 6\nThe ( 1 , 1 ) th element is: 16\nThe ( 1 , 2 ) th element is: 46\nThe ( 2 , 0 ) th element is: 11\nThe ( 2 , 1 ) th element is: 21\nThe ( 2 , 2 ) th element is: 51\n" ], [ "i=0\nwhile i<5:\n print( i, \"th turn\")\n i = i+1", "0 th turn\n1 th turn\n2 th turn\n3 th turn\n4 th turn\n" ], [ "for i in range(10):\n print(i)\n if i == 3:\n break", "0\n1\n2\n3\n" ], [ "import random as random\nfor i in range(10):\n r = random.uniform(1,10)\n if r<2 and r>0:\n print(\"It is smaller than 2 and greater than 1\",\"|\",r)\n elif r<4 and r>2:\n print(\"It is smaller than 4 and greater than 2\",\"|\",r)\n elif r<6 and r>4:\n print(\"It is smaller tha 6 and greater than 4\",\"|\",r)\n elif r<8 and r>6:\n print(\"It is smaller than 8 and greate than 6\",\"|\",r)\n elif r<10 and r>8:\n print(\"It is smaller than 10 and greater than 8\",\"|\",r)\n ", "It is smaller than 8 and greate than 6 | 7.776816896801124\nIt is smaller than 10 and greater than 8 | 8.900894365576073\nIt is smaller than 10 and greater than 8 | 8.303922292796068\nIt is smaller than 2 and greater than 1 | 1.9805026109354515\nIt is smaller than 4 and greater than 2 | 2.642984031672224\nIt is smaller than 4 and greater than 2 | 2.3577860630219387\nIt is smaller than 4 and greater than 2 | 2.707988363790994\nIt is smaller tha 6 and greater than 4 | 4.950782515186212\nIt is smaller tha 6 and greater than 4 | 4.474577783554499\nIt is smaller than 8 and greate than 6 | 6.260661038808822\n" ], [ "s = 0\nfor i in range(1000+1):\n s = s+i\ns", "_____no_output_____" ], [ "s = 0\nLE = []\nfor i in range(1001):\n if i%2 ==0:\n LE.append(i)\n s= s+i\ns, sum(LE)", "_____no_output_____" ] ] ]

[ "markdown", "code" ]

[ [ "markdown", "markdown", "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import json\nimport os\nimport sys\nimport fnmatch\n\n\nimport numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\n", "_____no_output_____" ], [ "df = pd.read_csv('big_stat.csv')#pd.read_csv('BBGM_League_7_all_seasons_Average_Stats.csv')\ndf = df[(df.G*df.MP > 100)]\ndf.shape", "_____no_output_____" ], [ "y = df.iloc[:,-15:]\nX = df.iloc[:,11:-17]\ny = y[(X['AST%'] >0) & (X['AST%'] < 100)]\nX = X[(X['AST%'] >0) & (X['AST%'] < 100)]\nX['MP'] = df.MP\nX['Hgt'] = df['Hgt']\ndf.columns\n", "_____no_output_____" ], [ "stat_list = ['FG','FGA','3P',\"3PA\",'FT','FTA',\\\n 'ORB','DRB','TRB','AST','TOV','STL',\"Blk\",\\\n 'PF','PTS']\nfor name in stat_list:\n den = np.maximum(1,df.MP)\n #if name in ['OWS','DWS']:\n # den = den*df['G']\n #X[name] = df[name]\n X[name + 'p36'] = 36* df[name]/den\n X[name + 'p100'] = X[name + 'p36']*4/3\n\nX['3PtP'] = (2/(1+np.exp(-X['3PAp100']))-1)*X['3P%']/100\nX['Creation'] = X['ASTp100']*0.1843+(X['PTSp100']+X['TOVp100'])*0.0969-2.3021*X['3PtP']+0.0582*(X['ASTp100']*(X['PTSp100']+X['TOVp100'])*X['3PtP'] )-1.1942\nX['Load'] = (X['ASTp100']-(0.38*X['Creation'])*0.75)+X['FGAp100']+X['FTAp100']*0.44+X['Creation']+X['TOVp100']\nX['cTOV'] = X['TOVp100']/X['Load']\nX['DPM'] = X['Blkp100']*0.802+X['DRBp100']*0.42-4.7-0.07551*X['PFp100']+1.597019*X['STLp100']-0.26385*X['TOVp100']\nX['OPM'] = -8.57647+0.6111*X['PTSp100']-0.33918*(0.44*X['FTAp100']+X['FGAp100'])+0.440814*X['FTAp100']+0.379745*X['3PAp100']+0.634044*X['ASTp100']+0.77827*X['ORBp100']-1.08855*X['TOVp100']+0.26262*X['STLp100']\nX['BPM'] = X['OPM'] + X['DPM']\nX['Age'] = df['Age']\n\n#X['PassP'] = ((X['ASTp100']-(0.38*X['Creation']))*0.752+ X['Creation'] + X['TOVp100']) ** 0.67\n#'OPM','DPM','cTOV','Load'#stat_list[:-2]+\nX = X[[_ for _ in X.columns if '%A' in _ or _[-1]=='r' or 'FGP' in _ or 'p36' in _ or _ in (['OPM','BPM','DPM','Creation','cTOV','Load','Age','MP'])]]\n\n\nreplacement_filter = (df.Salary > 0.5) & (df.Salary < 1.0)\nreplacement_player_mean_bs = X[replacement_filter].mean()\nreplacement_player_std_bs = X[replacement_filter].std()\nreplacement_player_cov_bs = X[replacement_filter].cov()\n\nreplacement_player_mean_r = y[replacement_filter].mean()\nreplacement_player_std_r = y[replacement_filter].std()\nreplacement_player_cov_r = y[replacement_filter].cov()\n\n\nreplacement_player_mean_r\n\nX.columns\n", "_____no_output_____" ], [ "np.round(replacement_player_std_r).astype(np.int)\n#_ = plt.hist(X['OWSp36'],150)", "_____no_output_____" ], [ "from sklearn import neural_network\nfrom sklearn import linear_model\nfrom sklearn import preprocessing\nfrom sklearn import feature_selection\nfrom sklearn import multioutput\nfrom sklearn import ensemble\nfrom sklearn import svm", "_____no_output_____" ], [ "fexp = preprocessing.PolynomialFeatures(degree=2,interaction_only=True)\nscalerX = preprocessing.RobustScaler()\nscalery = preprocessing.StandardScaler()\nprescale_X = scalerX.fit_transform(X)\nprescale_y = scalery.fit_transform(y)\nprescale_X = fexp.fit_transform(prescale_X)", "_____no_output_____" ], [ "trials = 1\nts = []\nfor i in range(trials):\n #clf = neural_network.MLPRegressor((36,5,24,36),'tanh',solver='adam',max_iter=1000)\n #clf = neural_network.MLPRegressor((),'identity',solver='lbfgs',alpha=5e2,tol=1e-9)\n #clf = multioutput.MultiOutputRegressor(linear_model.SGDRegressor(penalty='l2',alpha=5e2,eta0=1e-6,tol=1e-12,max_iter=50,verbose=True))\n clf = multioutput.MultiOutputRegressor(linear_model.ElasticNet(alpha=5e-3))\n #clf = ensemble.ExtraTreesRegressor(8,criterion='mae',max_depth=3,verbose=1)\n #clf = multioutput.MultiOutputRegressor(svm.SVR())\n clf.fit(prescale_X,prescale_y)\n yt = scalery.inverse_transform(clf.predict(prescale_X))\n err = np.linalg.norm(yt-y)\n ts.append((err,clf))", "_____no_output_____" ], [ "#np.array([est.alpha_ for est in clf.estimators_]).mean()\n#0.007250734067011631-7e-3", "_____no_output_____" ], [ "ts = sorted(ts)[::1] # why not the biggest error\nprint(ts[0][0])\nclf = ts[0][1]", "_____no_output_____" ], [ "col_names = X.columns\ncol_names = fexp.get_feature_names(X.columns)\n\nfor i,c in enumerate(y.columns):\n coeffs = clf.estimators_[i].coef_ \n v = np.argsort(abs(coeffs))[::-1]\n print(c)\n coeffs2 = [(coeffs[i2],col_names[i2]) for i2 in v[:10]]\n #for v,n in sorted(coeffs2,reverse=True):\n # print('{:.2f} * {} + '.format(v,n),end='')\n print('| Variable | Coeff |')\n print('|----------|-------|')\n for v,n in sorted(coeffs2,reverse=True):\n print('|{:25s}|{:.2f}|'.format(n,v))\n #for v,n in sorted(coeffs2,reverse=True):\n # print('\\t{:25s}\\t{:.2f}'.format(n,v))\n", "_____no_output_____" ], [ "GEN_YEAR = 2019\ntyear = [GEN_YEAR]\n#if tyear[0] < 1980:\n# print(\"MY PARSING OF THE TABLES IS WRONG WITHOUT the 2PA/3PA TRACKS\")\n# raise\n\nCURRENT_YEAR = 2019\nall_tables = {}\nfor ty in tyear:\n all_tables[ty] = np.load('wnba_{}.pkl'.format(ty))\nteams = all_tables[tyear[0]].keys()\n", "_____no_output_____" ], [ "df.MP.max()", "_____no_output_____" ], [ "player_stats = {k:{} for k in tyear}\ntable_columns = {}\nrosters = {}\nfor ty in tyear:\n tables = all_tables[ty]\n\n for team in tables:\n team_tables = tables[team]\n for table_name in team_tables:\n if table_name in ['draft-rights','team_and_opponent','conf','name','logo']:\n continue\n table = team_tables[table_name].fillna(0)\n #print(table_name)\n #print(table.index)\n for row in table.itertuples():\n name = row[0]\n name = name.replace('\\xa0\\xa0',' ').replace('.','')\n if name == 'Team Totals':\n continue\n nsplit = name.split(' ')\n if nsplit[-1] in ['Jr.','Sr.','I','II','III',\"IV\",'(TW)']:\n name = ' '.join(nsplit[:-1])\n rosters[name] = team\n player_table = player_stats[ty].get(name,{})\n player_row = player_table.get(table_name,[])\n player_row = player_row + [row]\n\n player_table[table_name] = player_row\n player_stats[ty][name] = player_table\n #if name == 'Dennis Smith Jr.' or name == 'Luka Doncic':\n # print(player_stats[ty][name],team)\n table_columns[table_name] = table.columns", "_____no_output_____" ], [ "table_mask = {}\nfor table in table_columns:\n table_mask[table] = [_.strip() !='' for _ in table_columns[table] ]\n table_columns[table] = [_ for _ in table_columns[table] if _.strip() != '']\n#for player in player_stats:\n# for table_in in player_stats[player]:\n# if 'on_off' in table_in or 'salaries' in table_in:\n# continue\n# if len(player_stats[player][table_in]) > 1:\n# pass\n #print(player,table_in,'MP' in player_stats[player][table_in][0]._fields)\n #print(player_stats[player][table_in][0])", "_____no_output_____" ], [ "# add playoff data to normal data\nif False:\n for ty in tyear:\n for player in player_stats[ty]:\n for table_in in player_stats[ty][player]:\n tableN = table_in.split('_')\n tableS = '_'.join(tableN[1:])\n if 'playoffs'==tableN[0] and not table_in in ['playoffs_pbp']:\n #print(table_in)\n if tableS in player_stats[ty][player]:\n player_stats[ty][player][tableS] += player_stats[ty][player][table_in]", "_____no_output_____" ], [ "for ty in tyear:\n for player in player_stats[ty]:\n for tt in player_stats[ty][player]:\n if tt in ['team_stats','team_stats_conf']:\n continue\n new_rows = []\n for tablet in player_stats[ty][player][tt]:\n vector = [_ if _ != '' else '0.0' for _ in tablet[1:]]\n vector = [(float(_.replace('%',''))/100 if type(_) == str and'%' in _ else _) for _ in vector]\n if 'on_off' in tt:\n vector = vector[1:]\n if 'contracts' in tt:\n vector = vector[1:-2]\n if tt in ['salaries2','contracts']:\n vector = [_.replace(',','').replace('$','') for _ in vector]\n try:\n v2 = np.array(vector).astype(np.float)\n except:\n v2 = vector\n new_rows.append(vector)\n a = np.array(new_rows)\n\n if 'MP' in table_columns[tt] and not tt in ['pbp','on_off','on_off_p']:\n try:\n a = a.astype(np.float)\n except:\n a = list(a)\n a[0] = np.array([float(_) for _ in a[0]])\n a[1] = np.array([float(_) for _ in a[1]])\n a = np.array(a)\n try:\n mins = a[:,table_columns[tt].index('MP')].reshape((-1,1))\n new_rows = ((a.T @ mins)/mins.sum()).T\n a = new_rows\n except:\n print(tt,a.shape,player,a,mins)\n pass\n\n player_stats[ty][player][tt] = a\n", "_____no_output_____" ], [ "#player_stats[2019]['Brandon Clarke']\n#player_stats[2019]\nfrom unidecode import unidecode\nwnba_ros = pd.read_csv('wnba_rosters.csv')\nwnba_ros.columns\nunidecode\n\n\nwnba_roster = {}\nfor row in wnba_ros.iterrows():\n name = unidecode(row[1]['name'])\n wnba_roster[name] = dict(row[1])\n hgt = wnba_roster[name]['hgt']\n hgt2 = hgt.split('-')\n if hgt == \"Jun-00\":\n wnba_roster[name]['hgt'] = '6-0'\n else:\n hgt2[1] = str({'May':5,'Jun':6}[hgt2[1]])\n wnba_roster[name]['hgt'] = hgt2[1] + '-' + hgt2[0]\n if wnba_roster[name]['exp'] == 'R':\n wnba_roster[name]['exp'] =0 \n wnba_roster[name]['exp'] = int(wnba_roster[name]['exp'])\n wnba_roster[name]['pos'] = wnba_roster[name]['pos'].replace('-','')", "_____no_output_____" ], [ "yr = wnba_roster['Kelsey Plum']['dob'].split('-')[-1]\nif yr[0] > '5':\n yr = '19' + yr\nelse:\n yr = '20' + yr\nyr", "_____no_output_____" ], [ "player_vectors = []\nplayer_names = []\nplayer_years = []\nplayer_scales = []\nplayer_heights = []\nplayer_confs = []\nplayer_composites = []\ncompNames = ['CUsage','CPassing','CTurnovers','CRim','CLowPost','CMidRange','C3Point','CFreeThrow','CRebound','CSteal','CBlock','CFouling','CDrawFoul']\n\nfor ty in tyear:\n for name in player_stats[ty]:\n #name = 'Jimmy Butler'\n #name = 'Ben Simmons'\n stats = player_stats[ty][name]\n if not 'wnba_totals' in stats:\n if ty == GEN_YEAR and name in player_stats[tyear[-1]] and 'wnba_totals' in player_stats[tyear[-1]][name]:\n stats = player_stats[tyear[-1]][name]\n else:\n continue\n if not name in wnba_roster:\n print(name)\n continue\n \n ht = [int(_) for _ in wnba_roster[name]['hgt'].split('-')]\n hgt = ht[0]*12 + ht[1]\n yr = wnba_roster['Kelsey Plum']['dob'].split('-')[-1]\n if yr[0] > '5':\n yr = '19' + yr\n else:\n yr = '20' + yr\n \n if ty >= 1980: # post 3pt era\n if 'wnba_advanced' in stats:\n d = {\n 'AtRimFGP':np.maximum(stats['wnba_advanced'][0][4],stats['wnba_per_game'][0][8])*100, #ts\n 'LowPostFGP':np.minimum(stats['wnba_advanced'][0][4],stats['wnba_per_game'][0][8])*100, #12\n 'MidRangeFGP':stats['wnba_per_game'][0][8]*100*.8}\n else:\n continue\n d['Age']= CURRENT_YEAR - int(yr)\n\n #print(name)\n MP = np.maximum(1,np.nan_to_num(stats['wnba_totals'][0][2]))\n\n try:\n #for stat in X.columns:\n d.update({'Hgt': hgt,\n 'FG%':stats['wnba_totals'][0][table_columns['wnba_per_game'].index('FG%')]*100,\n 'FG': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('FG')],\n 'FGA': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('FGA')],\n '3P': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('3P')],\n '3PA': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('3PA')],\n 'FT': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('FT')],\n 'FTA': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('FTA')],\n 'ORB': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('ORB')],\n 'DRB': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('DRB')],\n 'TRB': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('TRB')],\n 'AST': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('AST')],\n 'STL': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('STL')],\n 'Blk': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('BLK')],\n 'TOV': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('TOV')],\n 'PF': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('PF')],\n 'PTS': stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('PTS')],\n 'OWS':stats['wnba_advanced'][0][table_columns['wnba_advanced'].index('OWS')],\n 'DWS':stats['wnba_advanced'][0][table_columns['wnba_advanced'].index('DWS')],\n '3P%':stats['wnba_totals'][0][table_columns['wnba_totals'].index('3P%')]*100,\n 'MP':stats['wnba_per_game'][0][table_columns['wnba_per_game'].index('MP')]*48/40,\n 'FT%':stats['wnba_totals'][0][table_columns['wnba_totals'].index('FT%')]*100,\n 'TS%':stats['wnba_advanced'][0][table_columns['wnba_advanced'].index('TS%')]*100,\n '3PAr':stats['wnba_advanced'][0][table_columns['wnba_advanced'].index('3PAr')]*100,\n 'FTr':stats['wnba_advanced'][0][table_columns['wnba_advanced'].index('FTr')]*100,\n 'USG%':stats['wnba_advanced'][0][table_columns['wnba_advanced'].index('USG%')],\n 'FGp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('FG')],\n 'FGAp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('FGA')],\n '3Pp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('3P')],\n '3PAp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('3PA')],\n 'FTp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('FT')],\n 'FTAp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('FTA')],\n 'ORBp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('ORB')],\n 'DRBp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('DRB')],\n 'TRBp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('TRB')],\n 'ASTp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('AST')],\n 'TOVp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('TOV')],\n 'STLp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('STL')],\n 'Blkp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('BLK')],\n 'PFp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('PF')],\n 'PTSp36':stats['wnba_per_minute'][0][table_columns['wnba_per_minute'].index('PTS')],\n 'PER':stats['wnba_advanced'][0][table_columns['wnba_advanced'].index('PER')],})\n for k in list(d.keys()):\n if 'p36' in k:\n d[k[:-3]+'p100'] = d[k]*4/3\n\n d['3PtP'] = (2/(1+np.exp(-d['3PAp100']))-1)*d['3P%']/100\n d['Creation'] = d['ASTp100']*0.1843+(d['PTSp100']+d['TOVp100'])*0.0969-2.3021*d['3PtP']+0.0582*(d['ASTp100']*(d['PTSp100']+d['TOVp100'])*d['3PtP'] )-1.1942\n d['Load'] = (d['ASTp100']-(0.38*d['Creation'])*0.75)+d['FGAp100']+d['FTAp100']*0.44+d['Creation']+d['TOVp100']\n d['cTOV'] = d['TOVp100']/d['Load']\n d['DPM'] = d['Blkp100']*0.802+d['DRBp100']*0.42-4.7-0.07551*d['PFp100']+1.597019*d['STLp100']-0.26385*d['TOVp100']\n d['OPM'] = -8.57647+0.6111*d['PTSp100']-0.33918*(0.44*d['FTAp100']+d['FGAp100'])+0.440814*d['FTAp100']+0.379745*d['3PAp100']+0.634044*d['ASTp100']+0.77827*d['ORBp100']-1.08855*d['TOVp100']+0.26262*d['STLp100']\n d['BPM'] = d['OPM']+d['DPM']\n\n\n #if np.isnan(mp_scale):\n # print(name,ty,MP)\n player_scales.append(MP)\n #d['PassP'] = ((d['ASTp100']-(0.38*d['Creation']))*0.752+ d['Creation'] + d['TOVp100']) ** 0.67\n #if name == 'Marvin Bagley':\n # print(MP)\n # pprint.pprint({k:v for k,v in d.items() if not (('36' in k) or ('100' in k))} )\n player_vectors.append([d[stat] for stat in X.columns])\n player_heights.append(d['Hgt'])\n player_names.append(name)\n player_years.append(ty)\n #player_confs.append(rosters_conf[name])\n except KeyError:\n raise\n pass # player\n\nX.columns", "_____no_output_____" ], [ "#len(player_stats[1964]['Jerry West']['per_game'][0]),len(player_stats[1966]['Jerry West']['per_game'][0]),name\n#d,player_names[-1],X.columns,len(player_stats[1952]['Andy Phillip']['per_game'][0])\n#len(player_stats[1975]['Bob McAdoo']['per_game'][0])#[0][3] \n#stat_list\nn =['G', 'GS', 'MP', 'FG', 'FGA', 'FG%', '2P', '2PA', '2P%', '3P', '3PA', '3P%', 'FT', 'FTA', 'FT%', 'TRB', 'AST', 'STL', 'BLK', 'TOV', 'PF', 'PTS']\n#player_stats[2019]['James Harden']['pbp'][0][12]#[14]\n#for i in range(len(player_stats[2018]['Marvin Bagley']['per_min'][0])):\n# print(n[i], i,player_stats[2018]['Marvin Bagley']['per_min'][0][i])\n", "_____no_output_____" ], [ "first_n = len([yr for yr in player_years if yr == tyear[0]])\ngen_FA = len(teams)*1\nfirst_n,len(teams),gen_FA\n", "_____no_output_____" ], [ "Xn = np.nan_to_num(np.array(player_vectors))\n", "_____no_output_____" ], [ "#Xn = np.nan_to_num(np.array(player_vectors))\n# tuned this to get roughly 8-12 players at 70 or above. Which seemed like normal for a league\nscalerX2 = preprocessing.RobustScaler(quantile_range=(30.0, 55.0))\nscalerX2.fit(Xn[:first_n])\n#scalerX2 = scalerX\nXn_s =scalerX2.transform(np.nan_to_num(Xn))\nXn_fs = fexp.transform(np.nan_to_num(Xn_s))\npredict = clf.predict(Xn_fs)\nratings = np.nan_to_num(scalery.inverse_transform(predict))\nHGT_PRED = 0.25\n\nratings[:,0] = HGT_PRED*np.array(player_heights) + (1-HGT_PRED)*np.maximum(ratings[:,0],np.array(player_heights))\n\n# if we want to scale players down based on minutes played to replacement level\nif True:\n c = np.array(player_scales).reshape((-1,1))\n c = np.tanh(c/50) # basically 1 by 100 MP\n\n ratings[:Xn.shape[0]] = ratings[:Xn.shape[0]]*c + (1-c)*np.repeat(np.array(replacement_player_mean_r).reshape((-1,1)),Xn.shape[0],1).T", "_____no_output_____" ], [ "#for n,v in zip(X.columns,Xn_s[player_names.index('Draymond Green')]):\n# print(n,v)", "_____no_output_____" ], [ "\n#ratings[:,0]\n#table_columns['per_minute']#,table_columns['advanced'].index('0-3')\n#for i,t in enumerate(player_stats[name]['advanced'][0]):\n# print(i,t)", "_____no_output_____" ], [ "#player_vectors[player_names.index('Joel Embiid')][list(X.columns).index('OWSp36')],X.columns[list(X.columns).index('OWSp36')]\n#player_stats[2019]['Joel Embiid']['advanced'][0][12],player_stats[2019]['Joel Embiid']['per_game'][0][3]", "_____no_output_____" ], [ "X.mean(0)-Xn.mean(0)", "_____no_output_____" ], [ "#for n,v in zip(X.columns,Xn.max(0)):\n# print(n,v)\nall_tables[tyear[0]]['LVA']['wnba_totals']", "_____no_output_____" ], [ "tables.keys()", "_____no_output_____" ], [ "trades = {\n 'Coates, Alaina':'MIN',\n 'Talbot, Stephanie':'MIN',\n 'Lavender, Jantel':\"CHI\",\n 'McGee-Stafford, Imani': 'DAL',\n 'Cambage, Liz': \"LVA\",\n 'Harrison, Isabelle': 'DAL',\n 'Jefferson, Moriah': 'DAL',\n}\ntrades2 = {}\nfor name in trades:\n norig = name\n nsplit = [_.rstrip().strip() for _ in name.split(',')]\n name = nsplit[1] +' '+ nsplit[0]\n trades2[name] = trades[norig]\n\nrosters = {}\n\nfor team in all_tables[tyear[0]]:\n for name in all_tables[tyear[0]][team]['wnba_totals'].index:\n name = name.replace('\\xa0\\xa0',' ').replace('.','')\n nsplit = name.split(' ')\n if nsplit[-1] in ['Jr.','Sr.','I','II','III',\"IV\",'(TW)','Jr','Sr']:\n name = ' '.join(nsplit[:-1])\n if name in trades2:\n if trades2[name] != team:\n continue\n rosters[name] = team\n #print(team,)\n#print(rosters)", "_____no_output_____" ], [ "base = json.load(open('wnba_teams.json'))\n\n\nCONFS = []\n\nCONFS.append({'did':0, 'cid':0, 'name':'Eastern'})\n\nCONFS.append({'did':1, 'cid':1, 'name':'Western'})\n\nbase['gameAttributes'] = [{'key':'numGamesPlayoffSeries', 'value':[1,1,5,5]}]\nbase['gameAttributes'].append({'key':'divs', 'value': CONFS})\nbase['gameAttributes'].append({'key':'numGames', 'value': 34})\n\n", "_____no_output_____" ], [ "base", "_____no_output_____" ], [ "tids = {}\nfor t in base['teams']:\n tids[t['abbrev']] = t['tid']\n\nbase['startingSeason'] = tyear[0]\n\ny_keys = [_.lower() for _ in y.columns]\n\ny_map = { 'hgt': 'hgt',\n 'stre': 'str',\n 'spd': 'spd',\n 'jmp': 'jmp',\n 'endu': 'end',\n 'ins': 'ins',\n 'dnk': 'dnk',\n 'ft': 'ft.1',\n 'fg': '2pt',\n 'tp': '3pt',\n 'diq': 'diq',\n 'oiq': 'oiq',\n 'drb': 'drb',\n 'pss': 'pss',\n 'reb': 'reb' }", "_____no_output_____" ], [ "if False:\n ri = 0\n _ = plt.hist(yt[:,ri],100,normed=True,alpha=0.5,label='predicted')\n _ = plt.hist(y.iloc[:,ri],100,normed=True,alpha=0.5,label='labels')\n _ = plt.hist(ratings[:,ri],100,normed=True,alpha=0.5,label='dataset')\n plt.xlim(-100,100)\n plt.legend()\nelif False:\n plt.figure(figsize=(7*3,6*3))\n\n for ri in range(len(X.columns)):\n #ri = 13\n #plt.figure()\n plt.subplot(6,7,ri+1)\n\n _ = plt.hist(Xn[:,ri],100,density=True,alpha=0.5,label='NBA')\n _ = plt.hist(X.iloc[:,ri],100,density=True,alpha=0.5,label='BBGM')\n plt.legend()\n plt.title(X.columns[ri])# + ' ' + str(ri))\n plt.tight_layout()\n #plt.xlim(-100,100)\nfrom collections import OrderedDict\n\ndef write_roman(num):\n\n roman = OrderedDict()\n roman[1000] = \"M\"\n roman[900] = \"CM\"\n roman[500] = \"D\"\n roman[400] = \"CD\"\n roman[100] = \"C\"\n roman[90] = \"XC\"\n roman[50] = \"L\"\n roman[40] = \"XL\"\n roman[10] = \"X\"\n roman[9] = \"IX\"\n roman[5] = \"V\"\n roman[4] = \"IV\"\n roman[1] = \"I\"\n\n def roman_num(num):\n for r in roman.keys():\n x, y = divmod(num, r)\n yield roman[r] * x\n num -= (r * x)\n if num <= 0:\n break\n\n return \"\".join([a for a in roman_num(num)])", "_____no_output_____" ], [ "if gen_FA > 0:\n player_names_f = player_names + [\"Free Agent{}\".format(write_roman(i)) for i in range(gen_FA)]\n player_years_f = player_years + [GEN_YEAR for i in range(gen_FA)]\n # scale them down, we don't want a bunch of amazing replacements\n MEAN_S = 0.95\n STD_S = 0.25\n rp_ratings = np.random.multivariate_normal(MEAN_S*replacement_player_mean_r,STD_S*replacement_player_cov_r,size=(gen_FA))\n ratings_f = np.vstack([ratings,rp_ratings])\nelse:\n player_names_f = player_names\n player_years_f = player_years\n ratings_f = ratings", "_____no_output_____" ], [ "import pprint\nimport copy\nplayers = []\npp = pprint.PrettyPrinter()\nfor i,name in enumerate(player_names_f):\n py = player_years_f[i]\n name = player_names_f[i]\n sname = name.split(' ')\n new_player = {}\n new_player['firstName'] = sname[0]\n new_player['lastName'] = ' '.join(sname[1:])\n year_gap = 0\n if name in player_stats[py]:\n tid = tids[rosters[name]]\n \n try:\n new_player['tid'] = tid\n ht = [int(_) for _ in wnba_roster[name]['hgt'].split('-')]\n hgt = ht[0]*12 + ht[1]\n yr = wnba_roster[name]['dob'].split('-')[-1]\n if yr[0] > '5':\n yr = '19' + yr\n else:\n yr = '20' + yr\n yr = int(yr)\n print\n years_exp = int(wnba_roster[name]['exp'])\n \n try:\n weight = int(wnba_roster[name]['weight'])\n except:\n weight = 180\n \n new_player['pos'] = wnba_roster[name]['pos']\n new_player['born'] = {'year':yr,'loc':''}\n new_player['weight'] = weight\n new_player['hgt'] = hgt\n new_player['draft'] = {'year':GEN_YEAR-years_exp ,\"round\": 0, \"pick\": 0, \"tid\": -1, \"originalTid\": -1,}\n except:\n print(name)\n raise\n continue\n elif i > Xn.shape[0]:\n #print(name)\n new_player['tid'] = int(-1)\n new_player['weight'] = int(np.random.normal(220,20))\n new_player['hgt'] = int(np.random.normal(6*12+6,3))\n new_player['born'] = {'year':int(GEN_YEAR-np.random.normal(32,2)),'loc':''}\n new_player['pos'] = \"GF\"\n #print(new_player)\n else:\n #print(name)\n continue\n #if name == 'Ben Simmons':\n # print(new_player)\n #print(year_gap,py,name)\n try:\n scale_rookie = 1.0\n sub_rookie = 0\n r_vec = {k: scale_rookie*ratings_f[i,y_keys.index(km)]+sub_rookie for k,km in y_map.items()}\n\n r_vec = {k: int(np.clip(v,0,100)) for k,v in r_vec.items()}\n\n new_player['ratings'] = [r_vec]\n #new_player['ratings']\n players.append(new_player)\n except:\n print(name)\n raise\n #if name in ['Luka Doncic']: #\"Trevon Duval\",'LeBron James'\n #print(py,exp_years)\n # pp.pprint(new_player)\n #pp.pprint(player_vectors[player_names.index(name)])", "_____no_output_____" ], [ "base['players'] = players\nwith open('wnba_roster_{}.json'.format(tyear[0]),'wt') as fp:\n json.dump(base,fp, sort_keys=True)", "_____no_output_____" ], [ "len(ratings),len(ratings_f),len(ratings)+gen_FA", "_____no_output_____" ], [ "#[(_['region'],_['tid']) for _ in base['teams']]", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "import pandas as pd\nimport numpy as np\n", "_____no_output_____" ], [ "#import original file\n#df = pd.read_csv(\"/home/hsuyeemon/work/Data/yearly_data/2016.csv\")", "_____no_output_____" ], [ "#df.head()", "_____no_output_____" ], [ "#df.shape", "_____no_output_____" ], [ "#dates = df[\"date\"]", "_____no_output_____" ], [ "#dates.shape", "_____no_output_____" ], [ "#df.dtypes", "_____no_output_____" ], [ "#convert object type to datetime type\n#df['date'] = pd.to_datetime(df['date'])", "_____no_output_____" ], [ "#remove time stamp\n#df['date'] = df['date'].dt.date", "_____no_output_____" ], [ "#df.to_csv(\"normalized_2016.csv\")", "_____no_output_____" ], [ "import pandas_profiling ", "_____no_output_____" ], [ "#update profilng report\ndf1 = pd.read_csv(\"normalized_2016.csv\") \ndf1\n", "_____no_output_____" ], [ "df1.drop(['Unnamed: 0','id','customerType','customerId'], axis=1, inplace=True)\n#profile = df1.profile_report(title='Pandas Profiling Report') \n#profile.to_file(output_file=\"profile2016.html\") \ndf1", "_____no_output_____" ], [ "df1.dropna(inplace=True)\ndf1", "_____no_output_____" ], [ "duplicateRowsDF2 =df1[df1.duplicated(subset=['date','saturation','storeId','productId','color'],keep=False)]", "_____no_output_____" ], [ "duplicateRowsDF2", "_____no_output_____" ], [ "#duplicateRowsDF2['date'] = pd.to_datetime(duplicateRowsDF2['date'])", "_____no_output_____" ], [ "#duplicateRowsDF2 = duplicateRowsDF2.sort_values(by='date')", "_____no_output_____" ], [ "#duplicateRowsDF2['date'] = duplicateRowsDF2['date'].dt.date", "_____no_output_____" ], [ "duplicateRowsDF2.to_csv(\"test2.csv\")", "_____no_output_____" ], [ "duplicateRowsDF2.head(10)", "_____no_output_____" ], [ "#update profilng report\ndf1 = pd.read_csv(\"normalized_2016.csv\") \ndf1\n", "_____no_output_____" ], [ "df1 = pd.read_csv(\"2016_prepared.csv\") ", "/home/hsuyeemon/anaconda3/envs/my_env/lib/python3.8/site-packages/IPython/core/interactiveshell.py:3062: DtypeWarning: Columns (3) have mixed types.Specify dtype option on import or set low_memory=False.\n has_raised = await self.run_ast_nodes(code_ast.body, cell_name,\n" ], [ "duplicateRowsDF2 =df1[df1.duplicated(['date','saturation','productId','color'],keep=False)]", "_____no_output_____" ], [ "duplicateRowsDF2.head(10)", "_____no_output_____" ], [ "df1", "_____no_output_____" ], [ "#for only one store storeId = 3930 and product Id = 051142505\nnew_csv = df1.loc[(df1['storeId'] == 3930.0) & (df1['productId'] == '051142505')]", "_____no_output_____" ], [ "new_csv.to_csv(\"2016_3930_051142505.csv\")", "_____no_output_____" ], [ "from matplotlib import pyplot\n# load dataset\ndataset = pd.read_csv('2016_3930_051142505.csv', header=0, index_col=0)\ndataset.drop(['storeId','productId'], axis=1, inplace=True)\nvalues = dataset.values", "_____no_output_____" ], [ "# specify columns to plot\ngroups = [0, 1, 2, 3]\ni = 1\n# plot each column\npyplot.figure()\npyplot.plot(values[:, group])\npyplot.title(dataset.columns[3], y=0.5, loc='right')\npyplot.show()", "_____no_output_____" ], [ "dateD =dataset[dataset.duplicated(['date'],keep=False)]", "_____no_output_____" ], [ "dateD.sort_values(by='date')", "_____no_output_____" ], [ "dateProduct =dataset[dataset.duplicated(['date','productId'],keep=False)]", "_____no_output_____" ], [ "dateProduct = dateProduct.sort_values(by=['date','productId'])", "_____no_output_____" ], [ "dateProduct.head(20)", "_____no_output_____" ], [ "dateProductColor =dataset[dataset.duplicated(['date','productId','color'],keep=False)]\ndateProductColor = dateProduct.sort_values(by=['date','productId','color'])", "_____no_output_____" ], [ "dateProductColor.head(10)", "_____no_output_____" ], [ "from matplotlib import pyplot\n# load dataset\ndataset = pd.read_csv('2016_3930_051142505.csv', header=0, index_col=0)\ndataset.drop(['storeId','productId','color','saturation'], axis=1, inplace=True)\nvalues = dataset.values\n\n# plot each column\npyplot.figure()\npyplot.plot(values[:, 1])\npyplot.title(dataset.columns[1], y=0.5, loc='right')\npyplot.show()", "_____no_output_____" ], [ "dataset", "_____no_output_____" ], [ "dup =dataset[dataset.duplicated(['date'],keep=False)]\ndup = dup.sort_values(by=['date'])", "_____no_output_____" ], [ "dup.head(10)", "_____no_output_____" ], [ "df = dataset.groupby(['date']).sum()", "_____no_output_____" ], [ "idx = pd.date_range('2016-01-01', '2016-12-31')\ndf = df.sort_values(by='date')\ns=df.iloc[:]\ntype(s)\n", "_____no_output_____" ], [ "\nprint(s)\n", " quantity\ndate \n2016-01-01 1\n2016-01-02 2\n2016-01-03 6\n2016-01-09 2\n2016-01-10 3\n2016-01-11 3\n2016-01-19 6\n2016-01-23 1\n2016-02-06 2\n2016-02-10 6\n2016-02-11 1\n2016-02-14 1\n2016-02-18 1\n2016-02-22 2\n2016-03-04 2\n2016-03-15 6\n2016-03-16 2\n2016-03-28 2\n2016-03-29 8\n2016-04-01 1\n2016-04-10 3\n2016-04-14 3\n2016-04-15 1\n2016-04-16 7\n2016-04-19 1\n2016-04-20 6\n2016-04-21 3\n2016-04-22 4\n2016-04-25 4\n2016-05-03 2\n2016-05-04 2\n2016-05-09 1\n2016-05-15 3\n2016-05-17 2\n2016-05-19 3\n2016-05-22 2\n2016-06-05 2\n2016-06-12 1\n2016-06-13 2\n2016-06-18 4\n2016-06-24 1\n2016-07-25 3\n2016-07-30 2\n2016-08-02 1\n" ], [ "s.index = pd.DatetimeIndex(s.index)\ns = s.reindex(idx, fill_value=0)", "_____no_output_____" ], [ "s", "_____no_output_____" ], [ "s.to_csv(\"demand.csv\")", "_____no_output_____" ], [ "# plot each column\npyplot.figure()\npyplot.plot(s[:, 1])\npyplot.title(s.columns[1], y=0.5, loc='right')\npyplot.show()", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "##### Copyright 2019 The TensorFlow Authors.", "_____no_output_____" ] ], [ [ "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n# you may not use this file except in compliance with the License.\n# You may obtain a copy of the License at\n#\n# https://www.apache.org/licenses/LICENSE-2.0\n#\n# Unless required by applicable law or agreed to in writing, software\n# distributed under the License is distributed on an \"AS IS\" BASIS,\n# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n# See the License for the specific language governing permissions and\n# limitations under the License.", "_____no_output_____" ] ], [ [ "# Transformer Chatbot", "_____no_output_____" ], [ "<table class=\"tfo-notebook-buttons\" align=\"left\">\n <td>\n <a target=\"_blank\" href=\"https://colab.research.google.com/github/tensorflow/examples/blob/master/community/en/transformer_chatbot.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n </td>\n <td>\n <a target=\"_blank\" href=\"https://colab.research.google.com/github/tensorflow/examples/blob/master/community/en/transformer_chatbot.ipynb\"><img src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\" />View source on GitHub</a>\n </td>\n</table>", "_____no_output_____" ], [ "This tutorial trains a <a href=\"https://arxiv.org/abs/1706.03762\" class=\"external\">Transformer model</a> to be a chatbot. This is an advanced example that assumes knowledge of [text generation](https://tensorflow.org/alpha/tutorials/text/text_generation), [attention](https://www.tensorflow.org/alpha/tutorials/text/nmt_with_attention) and [transformer](https://www.tensorflow.org/alpha/tutorials/text/transformer).\n\nThe core idea behind the Transformer model is *self-attention*—the ability to attend to different positions of the input sequence to compute a representation of that sequence. Transformer creates stacks of self-attention layers and is explained below in the sections *Scaled dot product attention* and *Multi-head attention*.\n\nNote: The model architecture is identical to the example in [Transformer model for language understanding](https://www.tensorflow.org/alpha/tutorials/text/transformer), and we demonstrate how to implement the same model in the Functional approach instead of Subclassing.", "_____no_output_____" ] ], [ [ "from __future__ import absolute_import, division, print_function, unicode_literals\n\ntry:\n # The %tensorflow_version magic only works in colab.\n %tensorflow_version 2.x\nexcept Exception:\n pass\nimport tensorflow as tf\ntf.random.set_seed(1234)\n\n!pip install tfds-nightly\nimport tensorflow_datasets as tfds\n\nimport os\nimport re\nimport numpy as np\n\nimport matplotlib.pyplot as plt\n", "Collecting tf-nightly-gpu-2.0-preview==2.0.0.dev20190520\n\u001b[?25l Downloading https://files.pythonhosted.org/packages/c9/c1/fcaf4f6873777da2cd3a7a8ac3c9648cef7c7413f13b8135521eb9b9804a/tf_nightly_gpu_2.0_preview-2.0.0.dev20190520-cp36-cp36m-manylinux1_x86_64.whl (349.0MB)\n\u001b[K |████████████████████████████████| 349.0MB 31kB/s \n\u001b[?25hRequirement already satisfied: tfds-nightly in /usr/local/lib/python3.6/dist-packages (1.0.2.dev201905140105)\nRequirement already satisfied: wheel>=0.26 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (0.33.4)\nRequirement already satisfied: gast>=0.2.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (0.2.2)\nRequirement already satisfied: termcolor>=1.1.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (1.1.0)\nRequirement already satisfied: six>=1.10.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (1.12.0)\nCollecting wrapt>=1.11.1 (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520)\n Downloading https://files.pythonhosted.org/packages/67/b2/0f71ca90b0ade7fad27e3d20327c996c6252a2ffe88f50a95bba7434eda9/wrapt-1.11.1.tar.gz\nRequirement already satisfied: numpy<2.0,>=1.14.5 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (1.16.3)\nRequirement already satisfied: protobuf>=3.6.1 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (3.7.1)\nRequirement already satisfied: astor>=0.6.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (0.7.1)\nRequirement already satisfied: grpcio>=1.8.6 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (1.15.0)\nCollecting tb-nightly<1.15.0a0,>=1.14.0a0 (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520)\n\u001b[?25l Downloading https://files.pythonhosted.org/packages/6f/99/4220b50dc87814988e969cc859c07d070423bea820bc24d16c2023057eb6/tb_nightly-1.14.0a20190520-py3-none-any.whl (3.1MB)\n\u001b[K |████████████████████████████████| 3.1MB 33.7MB/s \n\u001b[?25hCollecting google-pasta>=0.1.6 (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520)\n\u001b[?25l Downloading https://files.pythonhosted.org/packages/f9/68/a14620bfb042691f532dcde8576ff82ee82e4c003cdc0a3dbee5f289cee6/google_pasta-0.1.6-py3-none-any.whl (51kB)\n\u001b[K |████████████████████████████████| 61kB 27.4MB/s \n\u001b[?25hRequirement already satisfied: keras-applications>=1.0.6 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (1.0.7)\nRequirement already satisfied: keras-preprocessing>=1.0.5 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (1.0.9)\nRequirement already satisfied: absl-py>=0.7.0 in /usr/local/lib/python3.6/dist-packages (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (0.7.1)\nCollecting tensorflow-estimator-2.0-preview (from tf-nightly-gpu-2.0-preview==2.0.0.dev20190520)\n\u001b[?25l Downloading https://files.pythonhosted.org/packages/71/e7/779651eca277d48486ae03d007162d37c93449bc29358fbe748e13639734/tensorflow_estimator_2.0_preview-1.14.0.dev2019052000-py2.py3-none-any.whl (427kB)\n\u001b[K |████████████████████████████████| 430kB 51.7MB/s \n\u001b[?25hRequirement already satisfied: promise in /usr/local/lib/python3.6/dist-packages (from tfds-nightly) (2.2.1)\nRequirement already satisfied: future in /usr/local/lib/python3.6/dist-packages (from tfds-nightly) (0.16.0)\nRequirement already satisfied: dill in /usr/local/lib/python3.6/dist-packages (from tfds-nightly) (0.2.9)\nRequirement already satisfied: psutil in /usr/local/lib/python3.6/dist-packages (from tfds-nightly) (5.4.8)\nRequirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from tfds-nightly) (2.21.0)\nRequirement already satisfied: tensorflow-metadata in /usr/local/lib/python3.6/dist-packages (from tfds-nightly) (0.13.0)\nRequirement already satisfied: tqdm in /usr/local/lib/python3.6/dist-packages (from tfds-nightly) (4.28.1)\nRequirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from protobuf>=3.6.1->tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (41.0.1)\nRequirement already satisfied: werkzeug>=0.11.15 in /usr/local/lib/python3.6/dist-packages (from tb-nightly<1.15.0a0,>=1.14.0a0->tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (0.15.3)\nRequirement already satisfied: markdown>=2.6.8 in /usr/local/lib/python3.6/dist-packages (from tb-nightly<1.15.0a0,>=1.14.0a0->tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (3.1)\nRequirement already satisfied: h5py in /usr/local/lib/python3.6/dist-packages (from keras-applications>=1.0.6->tf-nightly-gpu-2.0-preview==2.0.0.dev20190520) (2.8.0)\nRequirement already satisfied: urllib3<1.25,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests->tfds-nightly) (1.24.3)\nRequirement already satisfied: chardet<3.1.0,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->tfds-nightly) (3.0.4)\nRequirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests->tfds-nightly) (2019.3.9)\nRequirement already satisfied: idna<2.9,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->tfds-nightly) (2.8)\nRequirement already satisfied: googleapis-common-protos in /usr/local/lib/python3.6/dist-packages (from tensorflow-metadata->tfds-nightly) (1.5.10)\nBuilding wheels for collected packages: wrapt\n Building wheel for wrapt (setup.py) ... \u001b[?25l\u001b[?25hdone\n Stored in directory: /root/.cache/pip/wheels/89/67/41/63cbf0f6ac0a6156588b9587be4db5565f8c6d8ccef98202fc\nSuccessfully built wrapt\n\u001b[31mERROR: thinc 6.12.1 has requirement wrapt<1.11.0,>=1.10.0, but you'll have wrapt 1.11.1 which is incompatible.\u001b[0m\nInstalling collected packages: wrapt, tb-nightly, google-pasta, tensorflow-estimator-2.0-preview, tf-nightly-gpu-2.0-preview\n Found existing installation: wrapt 1.10.11\n Uninstalling wrapt-1.10.11:\n Successfully uninstalled wrapt-1.10.11\nSuccessfully installed google-pasta-0.1.6 tb-nightly-1.14.0a20190520 tensorflow-estimator-2.0-preview-1.14.0.dev2019052000 tf-nightly-gpu-2.0-preview-2.0.0.dev20190520 wrapt-1.11.1\n" ] ], [ [ "##Prepare Dataset", "_____no_output_____" ], [ "We will use the conversations in movies and TV shows provided by [Cornell Movie-Dialogs Corpus](https://www.cs.cornell.edu/~cristian/Cornell_Movie-Dialogs_Corpus.html), which contains more than 220 thousands conversational exchanges between more than 10k pairs of movie characters, as our dataset.\n\n`movie_conversations.txt` contains list of the conversation IDs and `movie_lines.text` contains the text of assoicated with each conversation ID. For further information regarding the dataset, please check the README file in the zip file.\n", "_____no_output_____" ] ], [ [ "path_to_zip = tf.keras.utils.get_file(\n 'cornell_movie_dialogs.zip',\n origin=\n 'http://www.cs.cornell.edu/~cristian/data/cornell_movie_dialogs_corpus.zip',\n extract=True)\n\npath_to_dataset = os.path.join(\n os.path.dirname(path_to_zip), \"cornell movie-dialogs corpus\")\n\npath_to_movie_lines = os.path.join(path_to_dataset, 'movie_lines.txt')\npath_to_movie_conversations = os.path.join(path_to_dataset,\n 'movie_conversations.txt')", "Downloading data from http://www.cs.cornell.edu/~cristian/data/cornell_movie_dialogs_corpus.zip\n9920512/9916637 [==============================] - 1s 0us/step\n" ] ], [ [ "### Load and preprocess data\n\nTo keep this example simple and fast, we are limiting the maximum number of training samples to`MAX_SAMPLES=25000` and the maximum length of the sentence to be `MAX_LENGTH=40`.\n\nWe preprocess our dataset in the following order:\n* Extract `MAX_SAMPLES` conversation pairs into list of `questions` and `answers.\n* Preprocess each sentence by removing special characters in each sentence.\n* Build tokenizer (map text to ID and ID to text) using [TensorFlow Datasets SubwordTextEncoder](https://www.tensorflow.org/datasets/api_docs/python/tfds/features/text/SubwordTextEncoder).\n* Tokenize each sentence and add `START_TOKEN` and `END_TOKEN` to indicate the start and end of each sentence.\n* Filter out sentence that has more than `MAX_LENGTH` tokens.\n* Pad tokenized sentences to `MAX_LENGTH`\n\n", "_____no_output_____" ] ], [ [ "# Maximum number of samples to preprocess\nMAX_SAMPLES = 50000\n\ndef preprocess_sentence(sentence):\n sentence = sentence.lower().strip()\n # creating a space between a word and the punctuation following it\n # eg: \"he is a boy.\" => \"he is a boy .\"\n sentence = re.sub(r\"([?.!,])\", r\" \\1 \", sentence)\n sentence = re.sub(r'[\" \"]+', \" \", sentence)\n # replacing everything with space except (a-z, A-Z, \".\", \"?\", \"!\", \",\")\n sentence = re.sub(r\"[^a-zA-Z?.!,]+\", \" \", sentence)\n sentence = sentence.strip()\n # adding a start and an end token to the sentence\n return sentence\n\n\ndef load_conversations():\n # dictionary of line id to text\n id2line = {}\n with open(path_to_movie_lines, errors='ignore') as file:\n lines = file.readlines()\n for line in lines:\n parts = line.replace('\\n', '').split(' +++$+++ ')\n id2line[parts[0]] = parts[4]\n\n inputs, outputs = [], []\n with open(path_to_movie_conversations, 'r') as file:\n lines = file.readlines()\n for line in lines:\n parts = line.replace('\\n', '').split(' +++$+++ ')\n # get conversation in a list of line ID\n conversation = [line[1:-1] for line in parts[3][1:-1].split(', ')]\n for i in range(len(conversation) - 1):\n inputs.append(preprocess_sentence(id2line[conversation[i]]))\n outputs.append(preprocess_sentence(id2line[conversation[i + 1]]))\n if len(inputs) >= MAX_SAMPLES:\n return inputs, outputs\n return inputs, outputs\n\n\nquestions, answers = load_conversations()", "_____no_output_____" ], [ "print('Sample question: {}'.format(questions[20]))\nprint('Sample answer: {}'.format(answers[20]))", "Sample question: i really , really , really wanna go , but i can t . not unless my sister goes .\nSample answer: i m workin on it . but she doesn t seem to be goin for him .\n" ], [ "# Build tokenizer using tfds for both questions and answers\ntokenizer = tfds.features.text.SubwordTextEncoder.build_from_corpus(\n questions + answers, target_vocab_size=2**13)\n\n# Define start and end token to indicate the start and end of a sentence\nSTART_TOKEN, END_TOKEN = [tokenizer.vocab_size], [tokenizer.vocab_size + 1]\n\n# Vocabulary size plus start and end token\nVOCAB_SIZE = tokenizer.vocab_size + 2", "_____no_output_____" ], [ "print('Tokenized sample question: {}'.format(tokenizer.encode(questions[20])))", "Tokenized sample question: [4, 281, 3, 281, 3, 143, 395, 176, 3, 42, 4, 38, 8191, 2, 37, 873, 27, 2031, 3096, 1]\n" ], [ "# Maximum sentence length\nMAX_LENGTH = 40\n\n\n# Tokenize, filter and pad sentences\ndef tokenize_and_filter(inputs, outputs):\n tokenized_inputs, tokenized_outputs = [], []\n \n for (sentence1, sentence2) in zip(inputs, outputs):\n # tokenize sentence\n sentence1 = START_TOKEN + tokenizer.encode(sentence1) + END_TOKEN\n sentence2 = START_TOKEN + tokenizer.encode(sentence2) + END_TOKEN\n # check tokenized sentence max length\n if len(sentence1) <= MAX_LENGTH and len(sentence2) <= MAX_LENGTH:\n tokenized_inputs.append(sentence1)\n tokenized_outputs.append(sentence2)\n \n # pad tokenized sentences\n tokenized_inputs = tf.keras.preprocessing.sequence.pad_sequences(\n tokenized_inputs, maxlen=MAX_LENGTH, padding='post')\n tokenized_outputs = tf.keras.preprocessing.sequence.pad_sequences(\n tokenized_outputs, maxlen=MAX_LENGTH, padding='post')\n \n return tokenized_inputs, tokenized_outputs\n\n\nquestions, answers = tokenize_and_filter(questions, answers)", "_____no_output_____" ], [ "print('Vocab size: {}'.format(VOCAB_SIZE))\nprint('Number of samples: {}'.format(len(questions)))", "Vocab size: 8333\nNumber of samples: 44095\n" ] ], [ [ "### Create `tf.data.Dataset`\n\nWe are going to use the [tf.data.Dataset API](https://www.tensorflow.org/api_docs/python/tf/data) to contruct our input pipline in order to utilize features like caching and prefetching to speed up the training process.\n\nThe transformer is an auto-regressive model: it makes predictions one part at a time, and uses its output so far to decide what to do next.\n\nDuring training this example uses teacher-forcing. Teacher forcing is passing the true output to the next time step regardless of what the model predicts at the current time step.\n\nAs the transformer predicts each word, self-attention allows it to look at the previous words in the input sequence to better predict the next word.\n\nTo prevent the model from peaking at the expected output the model uses a look-ahead mask.\n\nTarget is divided into `decoder_inputs` which padded as an input to the decoder and `cropped_targets` for calculating our loss and accuracy.", "_____no_output_____" ] ], [ [ "BATCH_SIZE = 64\nBUFFER_SIZE = 20000\n\n# decoder inputs use the previous target as input\n# remove START_TOKEN from targets\ndataset = tf.data.Dataset.from_tensor_slices((\n {\n 'inputs': questions,\n 'dec_inputs': answers[:, :-1]\n },\n {\n 'outputs': answers[:, 1:]\n },\n))\n\ndataset = dataset.cache()\ndataset = dataset.shuffle(BUFFER_SIZE)\ndataset = dataset.batch(BATCH_SIZE)\ndataset = dataset.prefetch(tf.data.experimental.AUTOTUNE)", "WARNING: Logging before flag parsing goes to stderr.\nW0521 04:24:35.050999 140098170382208 deprecation.py:323] From /usr/local/lib/python3.6/dist-packages/tensorflow/python/data/util/random_seed.py:58: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\nInstructions for updating:\nUse tf.where in 2.0, which has the same broadcast rule as np.where\n" ], [ "print(dataset)", "<PrefetchDataset shapes: ({inputs: (None, 40), dec_inputs: (None, 39)}, {outputs: (None, 39)}), types: ({inputs: tf.int32, dec_inputs: tf.int32}, {outputs: tf.int32})>\n" ] ], [ [ "## Attention\n\n", "_____no_output_____" ], [ "### Scaled dot product Attention\n\nThe scaled dot-product attention function used by the transformer takes three inputs: Q (query), K (key), V (value). The equation used to calculate the attention weights is:\n\n$$\\Large{Attention(Q, K, V) = softmax_k(\\frac{QK^T}{\\sqrt{d_k}}) V} $$\n\nAs the softmax normalization is done on the `key`, its values decide the amount of importance given to the `query`.\n\nThe output represents the multiplication of the attention weights and the `value` vector. This ensures that the words we want to focus on are kept as is and the irrelevant words are flushed out.\n\nThe dot-product attention is scaled by a factor of square root of the depth. This is done because for large values of depth, the dot product grows large in magnitude pushing the softmax function where it has small gradients resulting in a very hard softmax. \n\nFor example, consider that `query` and `key` have a mean of 0 and variance of 1. Their matrix multiplication will have a mean of 0 and variance of `dk`. Hence, *square root of `dk`* is used for scaling (and not any other number) because the matmul of `query` and `key` should have a mean of 0 and variance of 1, so that we get a gentler softmax.\n\nThe mask is multiplied with *-1e9 (close to negative infinity).* This is done because the mask is summed with the scaled matrix multiplication of `query` and `key` and is applied immediately before a softmax. The goal is to zero out these cells, and large negative inputs to softmax are near zero in the output.", "_____no_output_____" ] ], [ [ "def scaled_dot_product_attention(query, key, value, mask):\n \"\"\"Calculate the attention weights. \"\"\"\n matmul_qk = tf.matmul(query, key, transpose_b=True)\n\n # scale matmul_qk\n depth = tf.cast(tf.shape(key)[-1], tf.float32)\n logits = matmul_qk / tf.math.sqrt(depth)\n\n # add the mask to zero out padding tokens\n if mask is not None:\n logits += (mask * -1e9)\n\n # softmax is normalized on the last axis (seq_len_k)\n attention_weights = tf.nn.softmax(logits, axis=-1)\n\n output = tf.matmul(attention_weights, value)\n\n return output", "_____no_output_____" ] ], [ [ "### Multi-head attention\n\n<img src=\"https://www.tensorflow.org/images/tutorials/transformer/multi_head_attention.png\" width=\"500\" alt=\"multi-head attention\">\n\n\nMulti-head attention consists of four parts:\n* Linear layers and split into heads.\n* Scaled dot-product attention.\n* Concatenation of heads.\n* Final linear layer.\n\nEach multi-head attention block gets three inputs; Q (query), K (key), V (value). These are put through linear (Dense) layers and split up into multiple heads. \n\nThe `scaled_dot_product_attention` defined above is applied to each head (broadcasted for efficiency). An appropriate mask must be used in the attention step. The attention output for each head is then concatenated (using `tf.transpose`, and `tf.reshape`) and put through a final `Dense` layer.\n\nInstead of one single attention head, `query`, `key`, and `value` are split into multiple heads because it allows the model to jointly attend to information at different positions from different representational spaces. After the split each head has a reduced dimensionality, so the total computation cost is the same as a single head attention with full dimensionality.", "_____no_output_____" ] ], [ [ "class MultiHeadAttention(tf.keras.layers.Layer):\n\n def __init__(self, d_model, num_heads, name=\"multi_head_attention\"):\n super(MultiHeadAttention, self).__init__(name=name)\n self.num_heads = num_heads\n self.d_model = d_model\n\n assert d_model % self.num_heads == 0\n\n self.depth = d_model // self.num_heads\n\n self.query_dense = tf.keras.layers.Dense(units=d_model)\n self.key_dense = tf.keras.layers.Dense(units=d_model)\n self.value_dense = tf.keras.layers.Dense(units=d_model)\n\n self.dense = tf.keras.layers.Dense(units=d_model)\n\n def split_heads(self, inputs, batch_size):\n inputs = tf.reshape(\n inputs, shape=(batch_size, -1, self.num_heads, self.depth))\n return tf.transpose(inputs, perm=[0, 2, 1, 3])\n\n def call(self, inputs):\n query, key, value, mask = inputs['query'], inputs['key'], inputs[\n 'value'], inputs['mask']\n batch_size = tf.shape(query)[0]\n\n # linear layers\n query = self.query_dense(query)\n key = self.key_dense(key)\n value = self.value_dense(value)\n\n # split heads\n query = self.split_heads(query, batch_size)\n key = self.split_heads(key, batch_size)\n value = self.split_heads(value, batch_size)\n\n # scaled dot-product attention\n scaled_attention = scaled_dot_product_attention(query, key, value, mask)\n\n scaled_attention = tf.transpose(scaled_attention, perm=[0, 2, 1, 3])\n\n # concatenation of heads\n concat_attention = tf.reshape(scaled_attention,\n (batch_size, -1, self.d_model))\n\n # final linear layer\n outputs = self.dense(concat_attention)\n\n return outputs", "_____no_output_____" ] ], [ [ "## Transformer", "_____no_output_____" ], [ "### Masking\n\n", "_____no_output_____" ], [ "`create_padding_mask` and `create_look_ahead` are helper functions to creating masks to mask out padded tokens, we are going to use these helper functions as `tf.keras.layers.Lambda` layers.\n\nMask all the pad tokens (value `0`) in the batch to ensure the model does not treat padding as input.", "_____no_output_____" ] ], [ [ "def create_padding_mask(x):\n mask = tf.cast(tf.math.equal(x, 0), tf.float32)\n # (batch_size, 1, 1, sequence length)\n return mask[:, tf.newaxis, tf.newaxis, :]", "_____no_output_____" ], [ "print(create_padding_mask(tf.constant([[1, 2, 0, 3, 0], [0, 0, 0, 4, 5]])))", "tf.Tensor(\n[[[[0. 0. 1. 0. 1.]]]\n\n\n [[[1. 1. 1. 0. 0.]]]], shape=(2, 1, 1, 5), dtype=float32)\n" ] ], [ [ "Look-ahead mask to mask the future tokens in a sequence.\nWe also mask out pad tokens.\n\ni.e. To predict the third word, only the first and second word will be used", "_____no_output_____" ] ], [ [ "def create_look_ahead_mask(x):\n seq_len = tf.shape(x)[1]\n look_ahead_mask = 1 - tf.linalg.band_part(tf.ones((seq_len, seq_len)), -1, 0)\n padding_mask = create_padding_mask(x)\n return tf.maximum(look_ahead_mask, padding_mask)", "_____no_output_____" ], [ "print(create_look_ahead_mask(tf.constant([[1, 2, 0, 4, 5]])))", "tf.Tensor(\n[[[[0. 1. 1. 1. 1.]\n [0. 0. 1. 1. 1.]\n [0. 0. 1. 1. 1.]\n [0. 0. 1. 0. 1.]\n [0. 0. 1. 0. 0.]]]], shape=(1, 1, 5, 5), dtype=float32)\n" ] ], [ [ "### Positional encoding\n\nSince this model doesn't contain any recurrence or convolution, positional encoding is added to give the model some information about the relative position of the words in the sentence. \n\nThe positional encoding vector is added to the embedding vector. Embeddings represent a token in a d-dimensional space where tokens with similar meaning will be closer to each other. But the embeddings do not encode the relative position of words in a sentence. So after adding the positional encoding, words will be closer to each other based on the *similarity of their meaning and their position in the sentence*, in the d-dimensional space.\n\nSee the notebook on [positional encoding](https://github.com/tensorflow/examples/blob/master/community/en/position_encoding.ipynb) to learn more about it. The formula for calculating the positional encoding is as follows:\n\n$$\\Large{PE_{(pos, 2i)} = sin(pos / 10000^{2i / d_{model}})} $$\n$$\\Large{PE_{(pos, 2i+1)} = cos(pos / 10000^{2i / d_{model}})} $$", "_____no_output_____" ] ], [ [ "class PositionalEncoding(tf.keras.layers.Layer):\n\n def __init__(self, position, d_model):\n super(PositionalEncoding, self).__init__()\n self.pos_encoding = self.positional_encoding(position, d_model)\n\n def get_angles(self, position, i, d_model):\n angles = 1 / tf.pow(10000, (2 * (i // 2)) / tf.cast(d_model, tf.float32))\n return position * angles\n\n def positional_encoding(self, position, d_model):\n angle_rads = self.get_angles(\n position=tf.range(position, dtype=tf.float32)[:, tf.newaxis],\n i=tf.range(d_model, dtype=tf.float32)[tf.newaxis, :],\n d_model=d_model)\n # apply sin to even index in the array\n sines = tf.math.sin(angle_rads[:, 0::2])\n # apply cos to odd index in the array\n cosines = tf.math.cos(angle_rads[:, 1::2])\n\n pos_encoding = tf.concat([sines, cosines], axis=-1)\n pos_encoding = pos_encoding[tf.newaxis, ...]\n return tf.cast(pos_encoding, tf.float32)\n\n def call(self, inputs):\n return inputs + self.pos_encoding[:, :tf.shape(inputs)[1], :]", "_____no_output_____" ], [ "sample_pos_encoding = PositionalEncoding(50, 512)\n\nplt.pcolormesh(sample_pos_encoding.pos_encoding.numpy()[0], cmap='RdBu')\nplt.xlabel('Depth')\nplt.xlim((0, 512))\nplt.ylabel('Position')\nplt.colorbar()\nplt.show()", "_____no_output_____" ] ], [ [ "### Encoder Layer\n\nEach encoder layer consists of sublayers:\n\n1. Multi-head attention (with padding mask) \n2. 2 dense layers followed by dropout\n\nEach of these sublayers has a residual connection around it followed by a layer normalization. Residual connections help in avoiding the vanishing gradient problem in deep networks.\n\nThe output of each sublayer is `LayerNorm(x + Sublayer(x))`. The normalization is done on the `d_model` (last) axis.", "_____no_output_____" ] ], [ [ "def encoder_layer(units, d_model, num_heads, dropout, name=\"encoder_layer\"):\n inputs = tf.keras.Input(shape=(None, d_model), name=\"inputs\")\n padding_mask = tf.keras.Input(shape=(1, 1, None), name=\"padding_mask\")\n\n attention = MultiHeadAttention(\n d_model, num_heads, name=\"attention\")({\n 'query': inputs,\n 'key': inputs,\n 'value': inputs,\n 'mask': padding_mask\n })\n attention = tf.keras.layers.Dropout(rate=dropout)(attention)\n attention = tf.keras.layers.LayerNormalization(\n epsilon=1e-6)(inputs + attention)\n\n outputs = tf.keras.layers.Dense(units=units, activation='relu')(attention)\n outputs = tf.keras.layers.Dense(units=d_model)(outputs)\n outputs = tf.keras.layers.Dropout(rate=dropout)(outputs)\n outputs = tf.keras.layers.LayerNormalization(\n epsilon=1e-6)(attention + outputs)\n\n return tf.keras.Model(\n inputs=[inputs, padding_mask], outputs=outputs, name=name)", "_____no_output_____" ], [ "sample_encoder_layer = encoder_layer(\n units=512,\n d_model=128,\n num_heads=4,\n dropout=0.3,\n name=\"sample_encoder_layer\")\n\ntf.keras.utils.plot_model(\n sample_encoder_layer, to_file='encoder_layer.png', show_shapes=True)", "_____no_output_____" ] ], [ [ "### Encoder\n\nThe Encoder consists of:\n1. Input Embedding\n2. Positional Encoding\n3. `num_layers` encoder layers\n\nThe input is put through an embedding which is summed with the positional encoding. The output of this summation is the input to the encoder layers. The output of the encoder is the input to the decoder.", "_____no_output_____" ] ], [ [ "def encoder(vocab_size,\n num_layers,\n units,\n d_model,\n num_heads,\n dropout,\n name=\"encoder\"):\n inputs = tf.keras.Input(shape=(None,), name=\"inputs\")\n padding_mask = tf.keras.Input(shape=(1, 1, None), name=\"padding_mask\")\n\n embeddings = tf.keras.layers.Embedding(vocab_size, d_model)(inputs)\n embeddings *= tf.math.sqrt(tf.cast(d_model, tf.float32))\n embeddings = PositionalEncoding(vocab_size, d_model)(embeddings)\n\n outputs = tf.keras.layers.Dropout(rate=dropout)(embeddings)\n\n for i in range(num_layers):\n outputs = encoder_layer(\n units=units,\n d_model=d_model,\n num_heads=num_heads,\n dropout=dropout,\n name=\"encoder_layer_{}\".format(i),\n )([outputs, padding_mask])\n\n return tf.keras.Model(\n inputs=[inputs, padding_mask], outputs=outputs, name=name)", "_____no_output_____" ], [ "sample_encoder = encoder(\n vocab_size=8192,\n num_layers=2,\n units=512,\n d_model=128,\n num_heads=4,\n dropout=0.3,\n name=\"sample_encoder\")\n\ntf.keras.utils.plot_model(\n sample_encoder, to_file='encoder.png', show_shapes=True)", "_____no_output_____" ] ], [ [ "### Decoder Layer\n\nEach decoder layer consists of sublayers:\n\n1. Masked multi-head attention (with look ahead mask and padding mask)\n2. Multi-head attention (with padding mask). `value` and `key` receive the *encoder output* as inputs. `query` receives the *output from the masked multi-head attention sublayer.*\n3. 2 dense layers followed by dropout\n\nEach of these sublayers has a residual connection around it followed by a layer normalization. The output of each sublayer is `LayerNorm(x + Sublayer(x))`. The normalization is done on the `d_model` (last) axis.\n\nAs `query` receives the output from decoder's first attention block, and `key` receives the encoder output, the attention weights represent the importance given to the decoder's input based on the encoder's output. In other words, the decoder predicts the next word by looking at the encoder output and self-attending to its own output. See the demonstration above in the scaled dot product attention section.", "_____no_output_____" ] ], [ [ "def decoder_layer(units, d_model, num_heads, dropout, name=\"decoder_layer\"):\n inputs = tf.keras.Input(shape=(None, d_model), name=\"inputs\")\n enc_outputs = tf.keras.Input(shape=(None, d_model), name=\"encoder_outputs\")\n look_ahead_mask = tf.keras.Input(\n shape=(1, None, None), name=\"look_ahead_mask\")\n padding_mask = tf.keras.Input(shape=(1, 1, None), name='padding_mask')\n\n attention1 = MultiHeadAttention(\n d_model, num_heads, name=\"attention_1\")(inputs={\n 'query': inputs,\n 'key': inputs,\n 'value': inputs,\n 'mask': look_ahead_mask\n })\n attention1 = tf.keras.layers.LayerNormalization(\n epsilon=1e-6)(attention1 + inputs)\n\n attention2 = MultiHeadAttention(\n d_model, num_heads, name=\"attention_2\")(inputs={\n 'query': attention1,\n 'key': enc_outputs,\n 'value': enc_outputs,\n 'mask': padding_mask\n })\n attention2 = tf.keras.layers.Dropout(rate=dropout)(attention2)\n attention2 = tf.keras.layers.LayerNormalization(\n epsilon=1e-6)(attention2 + attention1)\n\n outputs = tf.keras.layers.Dense(units=units, activation='relu')(attention2)\n outputs = tf.keras.layers.Dense(units=d_model)(outputs)\n outputs = tf.keras.layers.Dropout(rate=dropout)(outputs)\n outputs = tf.keras.layers.LayerNormalization(\n epsilon=1e-6)(outputs + attention2)\n\n return tf.keras.Model(\n inputs=[inputs, enc_outputs, look_ahead_mask, padding_mask],\n outputs=outputs,\n name=name)", "_____no_output_____" ], [ "sample_decoder_layer = decoder_layer(\n units=512,\n d_model=128,\n num_heads=4,\n dropout=0.3,\n name=\"sample_decoder_layer\")\n\ntf.keras.utils.plot_model(\n sample_decoder_layer, to_file='decoder_layer.png', show_shapes=True)", "_____no_output_____" ] ], [ [ "### Decoder\n\nThe Decoder consists of:\n1. Output Embedding\n2. Positional Encoding\n3. N decoder layers\n\nThe target is put through an embedding which is summed with the positional encoding. The output of this summation is the input to the decoder layers. The output of the decoder is the input to the final linear layer.", "_____no_output_____" ] ], [ [ "def decoder(vocab_size,\n num_layers,\n units,\n d_model,\n num_heads,\n dropout,\n name='decoder'):\n inputs = tf.keras.Input(shape=(None,), name='inputs')\n enc_outputs = tf.keras.Input(shape=(None, d_model), name='encoder_outputs')\n look_ahead_mask = tf.keras.Input(\n shape=(1, None, None), name='look_ahead_mask')\n padding_mask = tf.keras.Input(shape=(1, 1, None), name='padding_mask')\n \n embeddings = tf.keras.layers.Embedding(vocab_size, d_model)(inputs)\n embeddings *= tf.math.sqrt(tf.cast(d_model, tf.float32))\n embeddings = PositionalEncoding(vocab_size, d_model)(embeddings)\n\n outputs = tf.keras.layers.Dropout(rate=dropout)(embeddings)\n\n for i in range(num_layers):\n outputs = decoder_layer(\n units=units,\n d_model=d_model,\n num_heads=num_heads,\n dropout=dropout,\n name='decoder_layer_{}'.format(i),\n )(inputs=[outputs, enc_outputs, look_ahead_mask, padding_mask])\n\n return tf.keras.Model(\n inputs=[inputs, enc_outputs, look_ahead_mask, padding_mask],\n outputs=outputs,\n name=name)", "_____no_output_____" ], [ "sample_decoder = decoder(\n vocab_size=8192,\n num_layers=2,\n units=512,\n d_model=128,\n num_heads=4,\n dropout=0.3,\n name=\"sample_decoder\")\n\ntf.keras.utils.plot_model(\n sample_decoder, to_file='decoder.png', show_shapes=True)", "_____no_output_____" ] ], [ [ "### Transformer\n\nTransformer consists of the encoder, decoder and a final linear layer. The output of the decoder is the input to the linear layer and its output is returned.", "_____no_output_____" ] ], [ [ "def transformer(vocab_size,\n num_layers,\n units,\n d_model,\n num_heads,\n dropout,\n name=\"transformer\"):\n inputs = tf.keras.Input(shape=(None,), name=\"inputs\")\n dec_inputs = tf.keras.Input(shape=(None,), name=\"dec_inputs\")\n\n enc_padding_mask = tf.keras.layers.Lambda(\n create_padding_mask, output_shape=(1, 1, None),\n name='enc_padding_mask')(inputs)\n # mask the future tokens for decoder inputs at the 1st attention block\n look_ahead_mask = tf.keras.layers.Lambda(\n create_look_ahead_mask,\n output_shape=(1, None, None),\n name='look_ahead_mask')(dec_inputs)\n # mask the encoder outputs for the 2nd attention block\n dec_padding_mask = tf.keras.layers.Lambda(\n create_padding_mask, output_shape=(1, 1, None),\n name='dec_padding_mask')(inputs)\n\n enc_outputs = encoder(\n vocab_size=vocab_size,\n num_layers=num_layers,\n units=units,\n d_model=d_model,\n num_heads=num_heads,\n dropout=dropout,\n )(inputs=[inputs, enc_padding_mask])\n\n dec_outputs = decoder(\n vocab_size=vocab_size,\n num_layers=num_layers,\n units=units,\n d_model=d_model,\n num_heads=num_heads,\n dropout=dropout,\n )(inputs=[dec_inputs, enc_outputs, look_ahead_mask, dec_padding_mask])\n\n outputs = tf.keras.layers.Dense(units=vocab_size, name=\"outputs\")(dec_outputs)\n\n return tf.keras.Model(inputs=[inputs, dec_inputs], outputs=outputs, name=name)", "_____no_output_____" ], [ "sample_transformer = transformer(\n vocab_size=8192,\n num_layers=4,\n units=512,\n d_model=128,\n num_heads=4,\n dropout=0.3,\n name=\"sample_transformer\")\n\ntf.keras.utils.plot_model(\n sample_transformer, to_file='transformer.png', show_shapes=True)", "_____no_output_____" ] ], [ [ "## Train model", "_____no_output_____" ], [ "### Initialize model\n\nTo keep this example small and relatively fast, the values for *num_layers, d_model, and units* have been reduced. See the [paper](https://arxiv.org/abs/1706.03762) for all the other versions of the transformer.", "_____no_output_____" ] ], [ [ "tf.keras.backend.clear_session()\n\n# Hyper-parameters\nNUM_LAYERS = 2\nD_MODEL = 256\nNUM_HEADS = 8\nUNITS = 512\nDROPOUT = 0.1\n\nmodel = transformer(\n vocab_size=VOCAB_SIZE,\n num_layers=NUM_LAYERS,\n units=UNITS,\n d_model=D_MODEL,\n num_heads=NUM_HEADS,\n dropout=DROPOUT)", "_____no_output_____" ] ], [ [ "### Loss function\n\nSince the target sequences are padded, it is important to apply a padding mask when calculating the loss.", "_____no_output_____" ] ], [ [ "def loss_function(y_true, y_pred):\n y_true = tf.reshape(y_true, shape=(-1, MAX_LENGTH - 1))\n \n loss = tf.keras.losses.SparseCategoricalCrossentropy(\n from_logits=True, reduction='none')(y_true, y_pred)\n\n mask = tf.cast(tf.not_equal(y_true, 0), tf.float32)\n loss = tf.multiply(loss, mask)\n\n return tf.reduce_mean(loss)", "_____no_output_____" ] ], [ [ "### Custom learning rate\n\nUse the Adam optimizer with a custom learning rate scheduler according to the formula in the [paper](https://arxiv.org/abs/1706.03762).\n\n$$\\Large{lrate = d_{model}^{-0.5} * min(step{\\_}num^{-0.5}, step{\\_}num * warmup{\\_}steps^{-1.5})}$$", "_____no_output_____" ] ], [ [ "class CustomSchedule(tf.keras.optimizers.schedules.LearningRateSchedule):\n\n def __init__(self, d_model, warmup_steps=4000):\n super(CustomSchedule, self).__init__()\n\n self.d_model = d_model\n self.d_model = tf.cast(self.d_model, tf.float32)\n\n self.warmup_steps = warmup_steps\n\n def __call__(self, step):\n arg1 = tf.math.rsqrt(step)\n arg2 = step * (self.warmup_steps**-1.5)\n\n return tf.math.rsqrt(self.d_model) * tf.math.minimum(arg1, arg2)", "_____no_output_____" ], [ "sample_learning_rate = CustomSchedule(d_model=128)\n\nplt.plot(sample_learning_rate(tf.range(200000, dtype=tf.float32)))\nplt.ylabel(\"Learning Rate\")\nplt.xlabel(\"Train Step\")", "_____no_output_____" ] ], [ [ "### Compile Model\n", "_____no_output_____" ] ], [ [ "learning_rate = CustomSchedule(D_MODEL)\n\noptimizer = tf.keras.optimizers.Adam(\n learning_rate, beta_1=0.9, beta_2=0.98, epsilon=1e-9)\n\ndef accuracy(y_true, y_pred):\n # ensure labels have shape (batch_size, MAX_LENGTH - 1)\n y_true = tf.reshape(y_true, shape=(-1, MAX_LENGTH - 1))\n accuracy = tf.metrics.SparseCategoricalAccuracy()(y_true, y_pred)\n return accuracy\n\nmodel.compile(optimizer=optimizer, loss=loss_function, metrics=[accuracy])", "_____no_output_____" ] ], [ [ "### Fit model\n\nTrain our transformer by simply calling `model.fit()`", "_____no_output_____" ] ], [ [ "EPOCHS = 20\n\nmodel.fit(dataset, epochs=EPOCHS)", "Epoch 1/20\n689/689 [==============================] - 97s 141ms/step - loss: 2.1146 - accuracy: 0.0249\nEpoch 2/20\n689/689 [==============================] - 81s 118ms/step - loss: 1.5008 - accuracy: 0.0530\nEpoch 3/20\n689/689 [==============================] - 82s 119ms/step - loss: 1.3940 - accuracy: 0.0653\nEpoch 4/20\n689/689 [==============================] - 82s 118ms/step - loss: 1.3313 - accuracy: 0.0719\nEpoch 5/20\n689/689 [==============================] - 82s 119ms/step - loss: 1.2744 - accuracy: 0.0765\nEpoch 6/20\n689/689 [==============================] - 82s 119ms/step - loss: 1.2223 - accuracy: 0.0801\nEpoch 7/20\n689/689 [==============================] - 82s 118ms/step - loss: 1.1670 - accuracy: 0.0832\nEpoch 8/20\n689/689 [==============================] - 82s 119ms/step - loss: 1.1050 - accuracy: 0.0861\nEpoch 9/20\n689/689 [==============================] - 82s 119ms/step - loss: 1.0503 - accuracy: 0.0889\nEpoch 10/20\n689/689 [==============================] - 82s 120ms/step - loss: 1.0002 - accuracy: 0.0917\nEpoch 11/20\n689/689 [==============================] - 82s 118ms/step - loss: 0.9540 - accuracy: 0.0945\nEpoch 12/20\n689/689 [==============================] - 82s 119ms/step - loss: 0.9122 - accuracy: 0.0973\nEpoch 13/20\n689/689 [==============================] - 82s 118ms/step - loss: 0.8744 - accuracy: 0.1001\nEpoch 14/20\n689/689 [==============================] - 82s 119ms/step - loss: 0.8396 - accuracy: 0.1029\nEpoch 15/20\n689/689 [==============================] - 82s 119ms/step - loss: 0.8082 - accuracy: 0.1056\nEpoch 16/20\n689/689 [==============================] - 82s 119ms/step - loss: 0.7799 - accuracy: 0.1082\nEpoch 17/20\n689/689 [==============================] - 82s 118ms/step - loss: 0.7540 - accuracy: 0.1108\nEpoch 18/20\n689/689 [==============================] - 82s 119ms/step - loss: 0.7300 - accuracy: 0.1134\nEpoch 19/20\n689/689 [==============================] - 82s 119ms/step - loss: 0.7076 - accuracy: 0.1158\nEpoch 20/20\n689/689 [==============================] - 82s 118ms/step - loss: 0.6881 - accuracy: 0.1183\n" ] ], [ [ "## Evaluate and predict\n\nThe following steps are used for evaluation:\n\n* Apply the same preprocessing method we used to create our dataset for the input sentence.\n* Tokenize the input sentence and add `START_TOKEN` and `END_TOKEN`. \n* Calculate the padding masks and the look ahead masks.\n* The decoder then outputs the predictions by looking at the encoder output and its own output.\n* Select the last word and calculate the argmax of that.\n* Concatentate the predicted word to the decoder input as pass it to the decoder.\n* In this approach, the decoder predicts the next word based on the previous words it predicted.\n\nNote: The model used here has less capacity and trained on a subset of the full dataset, hence its performance can be further improved.", "_____no_output_____" ] ], [ [ "def evaluate(sentence):\n sentence = preprocess_sentence(sentence)\n\n sentence = tf.expand_dims(\n START_TOKEN + tokenizer.encode(sentence) + END_TOKEN, axis=0)\n\n output = tf.expand_dims(START_TOKEN, 0)\n\n for i in range(MAX_LENGTH):\n predictions = model(inputs=[sentence, output], training=False)\n\n # select the last word from the seq_len dimension\n predictions = predictions[:, -1:, :]\n predicted_id = tf.cast(tf.argmax(predictions, axis=-1), tf.int32)\n\n # return the result if the predicted_id is equal to the end token\n if tf.equal(predicted_id, END_TOKEN[0]):\n break\n\n # concatenated the predicted_id to the output which is given to the decoder\n # as its input.\n output = tf.concat([output, predicted_id], axis=-1)\n\n return tf.squeeze(output, axis=0)\n\n\ndef predict(sentence):\n prediction = evaluate(sentence)\n\n predicted_sentence = tokenizer.decode(\n [i for i in prediction if i < tokenizer.vocab_size])\n\n print('Input: {}'.format(sentence))\n print('Output: {}'.format(predicted_sentence))\n\n return predicted_sentence", "_____no_output_____" ] ], [ [ "Let's test our model!", "_____no_output_____" ] ], [ [ "output = predict('Where have you been?')", "Input: Where have you been?\nOutput: i m not gonna bring you here .\n" ], [ "output = predict(\"It's a trap\")", "Input: It's a trap\nOutput: no , it s not .\n" ], [ "# feed the model with its previous output\nsentence = 'I am not crazy, my mother had me tested.'\nfor _ in range(5):\n sentence = predict(sentence)\n print('')", "Input: I am not crazy, my mother had me tested.\nOutput: you re a good man , roy . that s a good man , roy , you re a little girl , that s a good man . you re a little girl .\n\nInput: you re a good man , roy . that s a good man , roy , you re a little girl , that s a good man . you re a little girl .\nOutput: i m glad you re not a drug addict .\n\nInput: i m glad you re not a drug addict .\nOutput: the inheritance .\n\nInput: the inheritance .\nOutput: no , i don t know what to do . i just like to tell you .\n\nInput: no , i don t know what to do . i just like to tell you .\nOutput: i m not sure . i m not gonna mention my name .\n\n" ] ], [ [ "## Summary\n\nHere we are, we have implemented a Transformer in TensorFlow 2.0 in around 500 lines of code.\n\nIn this tutorial, we focus on the two different approaches to implement complex models with Functional API and Model subclassing, and how to incorporate them.\n\nTry using a different dataset or hyper-parameters to train the Transformer! Thanks for reading.\n", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "##### Copyright 2020 The TensorFlow Authors.", "_____no_output_____" ] ], [ [ "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n# you may not use this file except in compliance with the License.\n# You may obtain a copy of the License at\n#\n# https://www.apache.org/licenses/LICENSE-2.0\n#\n# Unless required by applicable law or agreed to in writing, software\n# distributed under the License is distributed on an \"AS IS\" BASIS,\n# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n# See the License for the specific language governing permissions and\n# limitations under the License.", "_____no_output_____" ] ], [ [ "# Calculate gradients", "_____no_output_____" ], [ "<table class=\"tfo-notebook-buttons\" align=\"left\">\n <td>\n <a target=\"_blank\" href=\"https://www.tensorflow.org/quantum/tutorials/gradients\"><img src=\"https://www.tensorflow.org/images/tf_logo_32px.png\" />View on TensorFlow.org</a>\n </td>\n <td>\n <a target=\"_blank\" href=\"https://colab.research.google.com/github/tensorflow/quantum/blob/master/docs/tutorials/gradients.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />Run in Google Colab</a>\n </td>\n <td>\n <a target=\"_blank\" href=\"https://github.com/tensorflow/quantum/blob/master/docs/tutorials/gradients.ipynb\"><img src=\"https://www.tensorflow.org/images/GitHub-Mark-32px.png\" />View source on GitHub</a>\n </td>\n <td>\n <a href=\"https://storage.googleapis.com/tensorflow_docs/quantum/docs/tutorials/gradients.ipynb\"><img src=\"https://www.tensorflow.org/images/download_logo_32px.png\" />Download notebook</a>\n </td>\n</table>", "_____no_output_____" ], [ "This tutorial explores gradient calculation algorithms for the expectation values of quantum circuits.\n\nCalculating the gradient of the expectation value of a certain observable in a quantum circuit is an involved process. Expectation values of observables do not have the luxury of having analytic gradient formulas that are always easy to write down—unlike traditional machine learning transformations such as matrix multiplication or vector addition that have analytic gradient formulas which are easy to write down. As a result, there are different quantum gradient calculation methods that come in handy for different scenarios. This tutorial compares and contrasts two different differentiation schemes.", "_____no_output_____" ], [ "## Setup", "_____no_output_____" ] ], [ [ "!pip install tensorflow==2.3.1", "_____no_output_____" ] ], [ [ "Install TensorFlow Quantum:", "_____no_output_____" ] ], [ [ "!pip install tensorflow-quantum", "_____no_output_____" ] ], [ [ "Now import TensorFlow and the module dependencies:", "_____no_output_____" ] ], [ [ "import tensorflow as tf\nimport tensorflow_quantum as tfq\n\nimport cirq\nimport sympy\nimport numpy as np\n\n# visualization tools\n%matplotlib inline\nimport matplotlib.pyplot as plt\nfrom cirq.contrib.svg import SVGCircuit", "_____no_output_____" ] ], [ [ "## 1. Preliminary\n\nLet's make the notion of gradient calculation for quantum circuits a little more concrete. Suppose you have a parameterized circuit like this one:", "_____no_output_____" ] ], [ [ "qubit = cirq.GridQubit(0, 0)\nmy_circuit = cirq.Circuit(cirq.Y(qubit)**sympy.Symbol('alpha'))\nSVGCircuit(my_circuit)", "_____no_output_____" ] ], [ [ "Along with an observable:", "_____no_output_____" ] ], [ [ "pauli_x = cirq.X(qubit)\npauli_x", "_____no_output_____" ] ], [ [ "Looking at this operator you know that $⟨Y(\\alpha)| X | Y(\\alpha)⟩ = \\sin(\\pi \\alpha)$", "_____no_output_____" ] ], [ [ "def my_expectation(op, alpha):\n \"\"\"Compute ⟨Y(alpha)| `op` | Y(alpha)⟩\"\"\"\n params = {'alpha': alpha}\n sim = cirq.Simulator()\n final_state_vector = sim.simulate(my_circuit, params).final_state_vector\n return op.expectation_from_state_vector(final_state_vector, {qubit: 0}).real\n\n\nmy_alpha = 0.3\nprint(\"Expectation=\", my_expectation(pauli_x, my_alpha))\nprint(\"Sin Formula=\", np.sin(np.pi * my_alpha))", "_____no_output_____" ] ], [ [ " and if you define $f_{1}(\\alpha) = ⟨Y(\\alpha)| X | Y(\\alpha)⟩$ then $f_{1}^{'}(\\alpha) = \\pi \\cos(\\pi \\alpha)$. Let's check this:", "_____no_output_____" ] ], [ [ "def my_grad(obs, alpha, eps=0.01):\n grad = 0\n f_x = my_expectation(obs, alpha)\n f_x_prime = my_expectation(obs, alpha + eps)\n return ((f_x_prime - f_x) / eps).real\n\n\nprint('Finite difference:', my_grad(pauli_x, my_alpha))\nprint('Cosine formula: ', np.pi * np.cos(np.pi * my_alpha))", "_____no_output_____" ] ], [ [ "## 2. The need for a differentiator\n\nWith larger circuits, you won't always be so lucky to have a formula that precisely calculates the gradients of a given quantum circuit. In the event that a simple formula isn't enough to calculate the gradient, the `tfq.differentiators.Differentiator` class allows you to define algorithms for computing the gradients of your circuits. For instance you can recreate the above example in TensorFlow Quantum (TFQ) with:", "_____no_output_____" ] ], [ [ "expectation_calculation = tfq.layers.Expectation(\n differentiator=tfq.differentiators.ForwardDifference(grid_spacing=0.01))\n\nexpectation_calculation(my_circuit,\n operators=pauli_x,\n symbol_names=['alpha'],\n symbol_values=[[my_alpha]])", "_____no_output_____" ] ], [ [ "However, if you switch to estimating expectation based on sampling (what would happen on a true device) the values can change a little bit. This means you now have an imperfect estimate:", "_____no_output_____" ] ], [ [ "sampled_expectation_calculation = tfq.layers.SampledExpectation(\n differentiator=tfq.differentiators.ForwardDifference(grid_spacing=0.01))\n\nsampled_expectation_calculation(my_circuit,\n operators=pauli_x,\n repetitions=500,\n symbol_names=['alpha'],\n symbol_values=[[my_alpha]])", "_____no_output_____" ] ], [ [ "This can quickly compound into a serious accuracy problem when it comes to gradients:", "_____no_output_____" ] ], [ [ "# Make input_points = [batch_size, 1] array.\ninput_points = np.linspace(0, 5, 200)[:, np.newaxis].astype(np.float32)\nexact_outputs = expectation_calculation(my_circuit,\n operators=pauli_x,\n symbol_names=['alpha'],\n symbol_values=input_points)\nimperfect_outputs = sampled_expectation_calculation(my_circuit,\n operators=pauli_x,\n repetitions=500,\n symbol_names=['alpha'],\n symbol_values=input_points)\nplt.title('Forward Pass Values')\nplt.xlabel('$x$')\nplt.ylabel('$f(x)$')\nplt.plot(input_points, exact_outputs, label='Analytic')\nplt.plot(input_points, imperfect_outputs, label='Sampled')\nplt.legend()", "_____no_output_____" ], [ "# Gradients are a much different story.\nvalues_tensor = tf.convert_to_tensor(input_points)\n\nwith tf.GradientTape() as g:\n g.watch(values_tensor)\n exact_outputs = expectation_calculation(my_circuit,\n operators=pauli_x,\n symbol_names=['alpha'],\n symbol_values=values_tensor)\nanalytic_finite_diff_gradients = g.gradient(exact_outputs, values_tensor)\n\nwith tf.GradientTape() as g:\n g.watch(values_tensor)\n imperfect_outputs = sampled_expectation_calculation(\n my_circuit,\n operators=pauli_x,\n repetitions=500,\n symbol_names=['alpha'],\n symbol_values=values_tensor)\nsampled_finite_diff_gradients = g.gradient(imperfect_outputs, values_tensor)\n\nplt.title('Gradient Values')\nplt.xlabel('$x$')\nplt.ylabel('$f^{\\'}(x)$')\nplt.plot(input_points, analytic_finite_diff_gradients, label='Analytic')\nplt.plot(input_points, sampled_finite_diff_gradients, label='Sampled')\nplt.legend()", "_____no_output_____" ] ], [ [ "Here you can see that although the finite difference formula is fast to compute the gradients themselves in the analytical case, when it came to the sampling based methods it was far too noisy. More careful techniques must be used to ensure a good gradient can be calculated. Next you will look at a much slower technique that wouldn't be as well suited for analytical expectation gradient calculations, but does perform much better in the real-world sample based case:", "_____no_output_____" ] ], [ [ "# A smarter differentiation scheme.\ngradient_safe_sampled_expectation = tfq.layers.SampledExpectation(\n differentiator=tfq.differentiators.ParameterShift())\n\nwith tf.GradientTape() as g:\n g.watch(values_tensor)\n imperfect_outputs = gradient_safe_sampled_expectation(\n my_circuit,\n operators=pauli_x,\n repetitions=500,\n symbol_names=['alpha'],\n symbol_values=values_tensor)\n\nsampled_param_shift_gradients = g.gradient(imperfect_outputs, values_tensor)\n\nplt.title('Gradient Values')\nplt.xlabel('$x$')\nplt.ylabel('$f^{\\'}(x)$')\nplt.plot(input_points, analytic_finite_diff_gradients, label='Analytic')\nplt.plot(input_points, sampled_param_shift_gradients, label='Sampled')\nplt.legend()", "_____no_output_____" ] ], [ [ "From the above you can see that certain differentiators are best used for particular research scenarios. In general, the slower sample-based methods that are robust to device noise, etc., are great differentiators when testing or implementing algorithms in a more \"real world\" setting. Faster methods like finite difference are great for analytical calculations and you want higher throughput, but aren't yet concerned with the device viability of your algorithm.", "_____no_output_____" ], [ "## 3. Multiple observables\n\nLet's introduce a second observable and see how TensorFlow Quantum supports multiple observables for a single circuit.", "_____no_output_____" ] ], [ [ "pauli_z = cirq.Z(qubit)\npauli_z", "_____no_output_____" ] ], [ [ "If this observable is used with the same circuit as before, then you have $f_{2}(\\alpha) = ⟨Y(\\alpha)| Z | Y(\\alpha)⟩ = \\cos(\\pi \\alpha)$ and $f_{2}^{'}(\\alpha) = -\\pi \\sin(\\pi \\alpha)$. Perform a quick check:", "_____no_output_____" ] ], [ [ "test_value = 0.\n\nprint('Finite difference:', my_grad(pauli_z, test_value))\nprint('Sin formula: ', -np.pi * np.sin(np.pi * test_value))", "_____no_output_____" ] ], [ [ "It's a match (close enough).\n\nNow if you define $g(\\alpha) = f_{1}(\\alpha) + f_{2}(\\alpha)$ then $g'(\\alpha) = f_{1}^{'}(\\alpha) + f^{'}_{2}(\\alpha)$. Defining more than one observable in TensorFlow Quantum to use along with a circuit is equivalent to adding on more terms to $g$.\n\nThis means that the gradient of a particular symbol in a circuit is equal to the sum of the gradients with regards to each observable for that symbol applied to that circuit. This is compatible with TensorFlow gradient taking and backpropagation (where you give the sum of the gradients over all observables as the gradient for a particular symbol).", "_____no_output_____" ] ], [ [ "sum_of_outputs = tfq.layers.Expectation(\n differentiator=tfq.differentiators.ForwardDifference(grid_spacing=0.01))\n\nsum_of_outputs(my_circuit,\n operators=[pauli_x, pauli_z],\n symbol_names=['alpha'],\n symbol_values=[[test_value]])", "_____no_output_____" ] ], [ [ "Here you see the first entry is the expectation w.r.t Pauli X, and the second is the expectation w.r.t Pauli Z. Now when you take the gradient:", "_____no_output_____" ] ], [ [ "test_value_tensor = tf.convert_to_tensor([[test_value]])\n\nwith tf.GradientTape() as g:\n g.watch(test_value_tensor)\n outputs = sum_of_outputs(my_circuit,\n operators=[pauli_x, pauli_z],\n symbol_names=['alpha'],\n symbol_values=test_value_tensor)\n\nsum_of_gradients = g.gradient(outputs, test_value_tensor)\n\nprint(my_grad(pauli_x, test_value) + my_grad(pauli_z, test_value))\nprint(sum_of_gradients.numpy())", "_____no_output_____" ] ], [ [ "Here you have verified that the sum of the gradients for each observable is indeed the gradient of $\\alpha$. This behavior is supported by all TensorFlow Quantum differentiators and plays a crucial role in the compatibility with the rest of TensorFlow.", "_____no_output_____" ], [ "## 4. Advanced usage\nHere you will learn how to define your own custom differentiation routines for quantum circuits.\nAll differentiators that exist inside of TensorFlow Quantum subclass `tfq.differentiators.Differentiator`. A differentiator must implement `differentiate_analytic` and `differentiate_sampled`.\n\nThe following uses TensorFlow Quantum constructs to implement the closed form solution from the first part of this tutorial.", "_____no_output_____" ] ], [ [ "class MyDifferentiator(tfq.differentiators.Differentiator):\n \"\"\"A Toy differentiator for <Y^alpha | X |Y^alpha>.\"\"\"\n\n def __init__(self):\n pass\n\n @tf.function\n def get_gradient_circuits(self, programs, symbol_names, symbol_values):\n \"\"\"Return circuits to compute gradients for given forward pass circuits.\n \n When implementing a gradient, it is often useful to describe the\n intermediate computations in terms of transformed versions of the input\n circuits. The details are beyond the scope of this tutorial, but interested\n users should check out the differentiator implementations in the TFQ library\n for examples.\n \"\"\"\n raise NotImplementedError(\n \"Gradient circuits are not implemented in this tutorial.\")\n\n @tf.function\n def _compute_gradient(self, symbol_values):\n \"\"\"Compute the gradient based on symbol_values.\"\"\"\n\n # f(x) = sin(pi * x)\n # f'(x) = pi * cos(pi * x)\n return tf.cast(tf.cos(symbol_values * np.pi) * np.pi, tf.float32)\n\n @tf.function\n def differentiate_analytic(self, programs, symbol_names, symbol_values,\n pauli_sums, forward_pass_vals, grad):\n \"\"\"Specify how to differentiate a circuit with analytical expectation.\n\n This is called at graph runtime by TensorFlow. `differentiate_analytic`\n should calculate the gradient of a batch of circuits and return it\n formatted as indicated below. See\n `tfq.differentiators.ForwardDifference` for an example.\n\n Args:\n programs: `tf.Tensor` of strings with shape [batch_size] containing\n the string representations of the circuits to be executed.\n symbol_names: `tf.Tensor` of strings with shape [n_params], which\n is used to specify the order in which the values in\n `symbol_values` should be placed inside of the circuits in\n `programs`.\n symbol_values: `tf.Tensor` of real numbers with shape\n [batch_size, n_params] specifying parameter values to resolve\n into the circuits specified by programs, following the ordering\n dictated by `symbol_names`.\n pauli_sums: `tf.Tensor` of strings with shape [batch_size, n_ops]\n containing the string representation of the operators that will\n be used on all of the circuits in the expectation calculations.\n forward_pass_vals: `tf.Tensor` of real numbers with shape\n [batch_size, n_ops] containing the output of the forward pass\n through the op you are differentiating.\n grad: `tf.Tensor` of real numbers with shape [batch_size, n_ops]\n representing the gradient backpropagated to the output of the\n op you are differentiating through.\n\n Returns:\n A `tf.Tensor` with the same shape as `symbol_values` representing\n the gradient backpropagated to the `symbol_values` input of the op\n you are differentiating through.\n \"\"\"\n\n # Computing gradients just based off of symbol_values.\n return self._compute_gradient(symbol_values) * grad\n\n @tf.function\n def differentiate_sampled(self, programs, symbol_names, symbol_values,\n pauli_sums, num_samples, forward_pass_vals, grad):\n \"\"\"Specify how to differentiate a circuit with sampled expectation.\n\n This is called at graph runtime by TensorFlow. `differentiate_sampled`\n should calculate the gradient of a batch of circuits and return it\n formatted as indicated below. See\n `tfq.differentiators.ForwardDifference` for an example.\n\n Args:\n programs: `tf.Tensor` of strings with shape [batch_size] containing\n the string representations of the circuits to be executed.\n symbol_names: `tf.Tensor` of strings with shape [n_params], which\n is used to specify the order in which the values in\n `symbol_values` should be placed inside of the circuits in\n `programs`.\n symbol_values: `tf.Tensor` of real numbers with shape\n [batch_size, n_params] specifying parameter values to resolve\n into the circuits specified by programs, following the ordering\n dictated by `symbol_names`.\n pauli_sums: `tf.Tensor` of strings with shape [batch_size, n_ops]\n containing the string representation of the operators that will\n be used on all of the circuits in the expectation calculations.\n num_samples: `tf.Tensor` of positive integers representing the\n number of samples per term in each term of pauli_sums used\n during the forward pass.\n forward_pass_vals: `tf.Tensor` of real numbers with shape\n [batch_size, n_ops] containing the output of the forward pass\n through the op you are differentiating.\n grad: `tf.Tensor` of real numbers with shape [batch_size, n_ops]\n representing the gradient backpropagated to the output of the\n op you are differentiating through.\n\n Returns:\n A `tf.Tensor` with the same shape as `symbol_values` representing\n the gradient backpropagated to the `symbol_values` input of the op\n you are differentiating through.\n \"\"\"\n return self._compute_gradient(symbol_values) * grad", "_____no_output_____" ] ], [ [ "This new differentiator can now be used with existing `tfq.layer` objects:", "_____no_output_____" ] ], [ [ "custom_dif = MyDifferentiator()\ncustom_grad_expectation = tfq.layers.Expectation(differentiator=custom_dif)\n\n# Now let's get the gradients with finite diff.\nwith tf.GradientTape() as g:\n g.watch(values_tensor)\n exact_outputs = expectation_calculation(my_circuit,\n operators=[pauli_x],\n symbol_names=['alpha'],\n symbol_values=values_tensor)\n\nanalytic_finite_diff_gradients = g.gradient(exact_outputs, values_tensor)\n\n# Now let's get the gradients with custom diff.\nwith tf.GradientTape() as g:\n g.watch(values_tensor)\n my_outputs = custom_grad_expectation(my_circuit,\n operators=[pauli_x],\n symbol_names=['alpha'],\n symbol_values=values_tensor)\n\nmy_gradients = g.gradient(my_outputs, values_tensor)\n\nplt.subplot(1, 2, 1)\nplt.title('Exact Gradient')\nplt.plot(input_points, analytic_finite_diff_gradients.numpy())\nplt.xlabel('x')\nplt.ylabel('f(x)')\nplt.subplot(1, 2, 2)\nplt.title('My Gradient')\nplt.plot(input_points, my_gradients.numpy())\nplt.xlabel('x')", "_____no_output_____" ] ], [ [ "This new differentiator can now be used to generate differentiable ops.\n\nKey Point: A differentiator that has been previously attached to an op must be refreshed before attaching to a new op, because a differentiator may only be attached to one op at a time.", "_____no_output_____" ] ], [ [ "# Create a noisy sample based expectation op.\nexpectation_sampled = tfq.get_sampled_expectation_op(\n cirq.DensityMatrixSimulator(noise=cirq.depolarize(0.01)))\n\n# Make it differentiable with your differentiator:\n# Remember to refresh the differentiator before attaching the new op\ncustom_dif.refresh()\ndifferentiable_op = custom_dif.generate_differentiable_op(\n sampled_op=expectation_sampled)\n\n# Prep op inputs.\ncircuit_tensor = tfq.convert_to_tensor([my_circuit])\nop_tensor = tfq.convert_to_tensor([[pauli_x]])\nsingle_value = tf.convert_to_tensor([[my_alpha]])\nnum_samples_tensor = tf.convert_to_tensor([[1000]])\n\nwith tf.GradientTape() as g:\n g.watch(single_value)\n forward_output = differentiable_op(circuit_tensor, ['alpha'], single_value,\n op_tensor, num_samples_tensor)\n\nmy_gradients = g.gradient(forward_output, single_value)\n\nprint('---TFQ---')\nprint('Foward: ', forward_output.numpy())\nprint('Gradient:', my_gradients.numpy())\nprint('---Original---')\nprint('Forward: ', my_expectation(pauli_x, my_alpha))\nprint('Gradient:', my_grad(pauli_x, my_alpha))", "_____no_output_____" ] ], [ [ "Success: Now you can use all the differentiators that TensorFlow Quantum has to offer—and define your own.", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown" ], [ "code" ], [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "%tensorflow_version 2.x\nimport tensorflow as tf\nimport numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\n\nfrom sklearn.preprocessing import MinMaxScaler \nfrom sklearn.model_selection import train_test_split", "_____no_output_____" ], [ "# Download the dataset, source: https://www.openml.org/d/31\n!wget https://www.openml.org/data/get_csv/1586225/php0iVrYT.csv", "--2020-02-27 21:34:53-- https://www.openml.org/data/get_csv/1586225/php0iVrYT.csv\nResolving www.openml.org (www.openml.org)... 131.155.11.11\nConnecting to www.openml.org (www.openml.org)|131.155.11.11|:443... connected.\nHTTP request sent, awaiting response... 200 OK\nLength: unspecified [text/plain]\nSaving to: ‘php0iVrYT.csv.1’\n\nphp0iVrYT.csv.1 [ <=> ] 10.27K --.-KB/s in 0s \n\n2020-02-27 21:34:54 (98.6 MB/s) - ‘php0iVrYT.csv.1’ saved [10512]\n\n" ], [ "# Import the data\ndataset = pd.read_csv(\"php0iVrYT.csv\")", "_____no_output_____" ], [ "\"\"\"\n* V1: Recency - months since last donation\n* V2: Frequency - total number of donation\n* V3: Monetary - total blood donated in c.c.\n* V4: Time - months since first donation), and a binary variable representing whether he/she donated blood in March 2007 \n (1 stand for donating blood; 0 stands for not donating blood).\n\nThe target attribute is a binary variable representing whether he/she donated blood in March 2007 (2 stands for donating blood; 1 stands for not donating blood).\n\"\"\"\ncols = ['Recency', 'Frequency', 'Monetary', 'Time','Class']\ndataset.columns = cols", "_____no_output_____" ], [ "# Substract 1 from Class column to convert the classification to binary classification\ndataset.Class = dataset.Class - 1", "_____no_output_____" ], [ "dataset.describe()", "_____no_output_____" ], [ "X = dataset.iloc[:, :-1].values # features\nY = dataset.iloc[:, -1].values # target variable", "_____no_output_____" ], [ "# Split the data\nX_train, X_test, y_train, y_test = train_test_split(X, Y, test_size = .25)", "_____no_output_____" ], [ "# Scale features\nscaler = MinMaxScaler(feature_range = (0, 10))\nX_train, X_test = scaler.fit_transform(X_train), scaler.fit_transform(X_test)", "_____no_output_____" ], [ "X_train", "_____no_output_____" ], [ "# Train the model\nmodel = tf.keras.models.Sequential([\n tf.keras.layers.Flatten(input_shape = (X_train.shape[1],)),\n tf.keras.layers.Dense(256, activation = 'relu'),\n tf.keras.layers.Dropout(.4),\n tf.keras.layers.Dense(1, activation = 'sigmoid')\n])\n\n# compile\nopt = tf.keras.optimizers.Adamax(learning_rate = 0.01)\nmodel.compile(optimizer = opt,\n loss = 'binary_crossentropy',\n metrics = ['accuracy'])\n\n\n# fit the model\nr = model.fit(X_train, y_train, validation_data = (X_test, y_test), epochs = 250, verbose = 2)\n\n# plot the loss\nplt.plot(r.history['loss'], label = 'loss')\nplt.plot(r.history['val_loss'], label = 'val_loss')\nplt.legend()\nplt.show()\nplt.plot(r.history['accuracy'], label = 'accuracy')\nplt.plot(r.history['val_accuracy'], label = 'val_accuracy')\nplt.legend()\nplt.show()", "Train on 561 samples, validate on 187 samples\nEpoch 1/250\n561/561 - 0s - loss: 0.5884 - accuracy: 0.7291 - val_loss: 0.4005 - val_accuracy: 0.8342\nEpoch 2/250\n561/561 - 0s - loss: 0.5389 - accuracy: 0.7576 - val_loss: 0.3762 - val_accuracy: 0.8342\nEpoch 3/250\n561/561 - 0s - loss: 0.5134 - accuracy: 0.7754 - val_loss: 0.3761 - val_accuracy: 0.8289\nEpoch 4/250\n561/561 - 0s - loss: 0.5062 - accuracy: 0.7701 - val_loss: 0.3758 - val_accuracy: 0.8128\nEpoch 5/250\n561/561 - 0s - loss: 0.5207 - accuracy: 0.7469 - val_loss: 0.3759 - val_accuracy: 0.8342\nEpoch 6/250\n561/561 - 0s - loss: 0.4950 - accuracy: 0.7807 - val_loss: 0.3764 - val_accuracy: 0.8182\nEpoch 7/250\n561/561 - 0s - loss: 0.4908 - accuracy: 0.7594 - val_loss: 0.3771 - val_accuracy: 0.8342\nEpoch 8/250\n561/561 - 0s - loss: 0.5013 - accuracy: 0.7772 - val_loss: 0.3761 - val_accuracy: 0.8289\nEpoch 9/250\n561/561 - 0s - loss: 0.4950 - accuracy: 0.7665 - val_loss: 0.3758 - val_accuracy: 0.8342\nEpoch 10/250\n561/561 - 0s - loss: 0.4982 - accuracy: 0.7665 - val_loss: 0.3788 - val_accuracy: 0.8342\nEpoch 11/250\n561/561 - 0s - loss: 0.5100 - accuracy: 0.7772 - val_loss: 0.3750 - val_accuracy: 0.8396\nEpoch 12/250\n561/561 - 0s - loss: 0.4971 - accuracy: 0.7683 - val_loss: 0.3751 - val_accuracy: 0.8396\nEpoch 13/250\n561/561 - 0s - loss: 0.4896 - accuracy: 0.7701 - val_loss: 0.3737 - val_accuracy: 0.8289\nEpoch 14/250\n561/561 - 0s - loss: 0.4919 - accuracy: 0.7772 - val_loss: 0.3780 - val_accuracy: 0.8396\nEpoch 15/250\n561/561 - 0s - loss: 0.4917 - accuracy: 0.7701 - val_loss: 0.3744 - val_accuracy: 0.8128\nEpoch 16/250\n561/561 - 0s - loss: 0.4971 - accuracy: 0.7665 - val_loss: 0.3752 - val_accuracy: 0.8396\nEpoch 17/250\n561/561 - 0s - loss: 0.4894 - accuracy: 0.7701 - val_loss: 0.3759 - val_accuracy: 0.8449\nEpoch 18/250\n561/561 - 0s - loss: 0.4950 - accuracy: 0.7629 - val_loss: 0.3743 - val_accuracy: 0.8289\nEpoch 19/250\n561/561 - 0s - loss: 0.4918 - accuracy: 0.7825 - val_loss: 0.3780 - val_accuracy: 0.8396\nEpoch 20/250\n561/561 - 0s - loss: 0.4933 - accuracy: 0.7736 - val_loss: 0.3761 - val_accuracy: 0.8342\nEpoch 21/250\n561/561 - 0s - loss: 0.4948 - accuracy: 0.7825 - val_loss: 0.3757 - val_accuracy: 0.8449\nEpoch 22/250\n561/561 - 0s - loss: 0.4956 - accuracy: 0.7754 - val_loss: 0.3765 - val_accuracy: 0.8396\nEpoch 23/250\n561/561 - 0s - loss: 0.4852 - accuracy: 0.7861 - val_loss: 0.3763 - val_accuracy: 0.8396\nEpoch 24/250\n561/561 - 0s - loss: 0.5013 - accuracy: 0.7754 - val_loss: 0.3781 - val_accuracy: 0.8289\nEpoch 25/250\n561/561 - 0s - loss: 0.4866 - accuracy: 0.7825 - val_loss: 0.3754 - val_accuracy: 0.8342\nEpoch 26/250\n561/561 - 0s - loss: 0.4906 - accuracy: 0.7825 - val_loss: 0.3764 - val_accuracy: 0.8289\nEpoch 27/250\n561/561 - 0s - loss: 0.4872 - accuracy: 0.7772 - val_loss: 0.3772 - val_accuracy: 0.8289\nEpoch 28/250\n561/561 - 0s - loss: 0.4849 - accuracy: 0.7772 - val_loss: 0.3776 - val_accuracy: 0.8182\nEpoch 29/250\n561/561 - 0s - loss: 0.4816 - accuracy: 0.7861 - val_loss: 0.3773 - val_accuracy: 0.8289\nEpoch 30/250\n561/561 - 0s - loss: 0.4917 - accuracy: 0.7754 - val_loss: 0.3774 - val_accuracy: 0.8289\nEpoch 31/250\n561/561 - 0s - loss: 0.4932 - accuracy: 0.7718 - val_loss: 0.3786 - val_accuracy: 0.8235\nEpoch 32/250\n561/561 - 0s - loss: 0.4891 - accuracy: 0.7754 - val_loss: 0.3769 - val_accuracy: 0.8235\nEpoch 33/250\n561/561 - 0s - loss: 0.4810 - accuracy: 0.7825 - val_loss: 0.3768 - val_accuracy: 0.8182\nEpoch 34/250\n561/561 - 0s - loss: 0.4865 - accuracy: 0.7754 - val_loss: 0.3795 - val_accuracy: 0.8182\nEpoch 35/250\n561/561 - 0s - loss: 0.4907 - accuracy: 0.7897 - val_loss: 0.3783 - val_accuracy: 0.8128\nEpoch 36/250\n561/561 - 0s - loss: 0.4881 - accuracy: 0.7790 - val_loss: 0.3804 - val_accuracy: 0.8182\nEpoch 37/250\n561/561 - 0s - loss: 0.4845 - accuracy: 0.7914 - val_loss: 0.3776 - val_accuracy: 0.8075\nEpoch 38/250\n561/561 - 0s - loss: 0.4764 - accuracy: 0.7897 - val_loss: 0.3781 - val_accuracy: 0.8182\nEpoch 39/250\n561/561 - 0s - loss: 0.4784 - accuracy: 0.7914 - val_loss: 0.3805 - val_accuracy: 0.8289\nEpoch 40/250\n561/561 - 0s - loss: 0.4798 - accuracy: 0.7861 - val_loss: 0.3793 - val_accuracy: 0.8075\nEpoch 41/250\n561/561 - 0s - loss: 0.4851 - accuracy: 0.7772 - val_loss: 0.3809 - val_accuracy: 0.8235\nEpoch 42/250\n561/561 - 0s - loss: 0.4815 - accuracy: 0.7932 - val_loss: 0.3787 - val_accuracy: 0.8075\nEpoch 43/250\n561/561 - 0s - loss: 0.4888 - accuracy: 0.7932 - val_loss: 0.3828 - val_accuracy: 0.8235\nEpoch 44/250\n561/561 - 0s - loss: 0.4829 - accuracy: 0.7932 - val_loss: 0.3781 - val_accuracy: 0.8075\nEpoch 45/250\n561/561 - 0s - loss: 0.4877 - accuracy: 0.7914 - val_loss: 0.3803 - val_accuracy: 0.8128\nEpoch 46/250\n561/561 - 0s - loss: 0.4766 - accuracy: 0.7897 - val_loss: 0.3798 - val_accuracy: 0.8075\nEpoch 47/250\n561/561 - 0s - loss: 0.4878 - accuracy: 0.7843 - val_loss: 0.3817 - val_accuracy: 0.8075\nEpoch 48/250\n561/561 - 0s - loss: 0.4882 - accuracy: 0.7754 - val_loss: 0.3814 - val_accuracy: 0.8235\nEpoch 49/250\n561/561 - 0s - loss: 0.4775 - accuracy: 0.7879 - val_loss: 0.3798 - val_accuracy: 0.8075\nEpoch 50/250\n561/561 - 0s - loss: 0.4797 - accuracy: 0.7950 - val_loss: 0.3821 - val_accuracy: 0.8235\nEpoch 51/250\n561/561 - 0s - loss: 0.4781 - accuracy: 0.7897 - val_loss: 0.3805 - val_accuracy: 0.8075\nEpoch 52/250\n561/561 - 0s - loss: 0.4773 - accuracy: 0.7897 - val_loss: 0.3829 - val_accuracy: 0.8182\nEpoch 53/250\n561/561 - 0s - loss: 0.4770 - accuracy: 0.7807 - val_loss: 0.3810 - val_accuracy: 0.8182\nEpoch 54/250\n561/561 - 0s - loss: 0.4828 - accuracy: 0.7825 - val_loss: 0.3820 - val_accuracy: 0.8075\nEpoch 55/250\n561/561 - 0s - loss: 0.4821 - accuracy: 0.7897 - val_loss: 0.3816 - val_accuracy: 0.8021\nEpoch 56/250\n561/561 - 0s - loss: 0.4829 - accuracy: 0.7879 - val_loss: 0.3829 - val_accuracy: 0.8075\nEpoch 57/250\n561/561 - 0s - loss: 0.4786 - accuracy: 0.7897 - val_loss: 0.3818 - val_accuracy: 0.8075\nEpoch 58/250\n561/561 - 0s - loss: 0.4767 - accuracy: 0.7861 - val_loss: 0.3837 - val_accuracy: 0.8235\nEpoch 59/250\n561/561 - 0s - loss: 0.4833 - accuracy: 0.7914 - val_loss: 0.3825 - val_accuracy: 0.8075\nEpoch 60/250\n561/561 - 0s - loss: 0.4778 - accuracy: 0.7897 - val_loss: 0.3847 - val_accuracy: 0.8235\nEpoch 61/250\n561/561 - 0s - loss: 0.4753 - accuracy: 0.7861 - val_loss: 0.3843 - val_accuracy: 0.8235\nEpoch 62/250\n561/561 - 0s - loss: 0.4811 - accuracy: 0.7932 - val_loss: 0.3835 - val_accuracy: 0.8021\nEpoch 63/250\n561/561 - 0s - loss: 0.4761 - accuracy: 0.7825 - val_loss: 0.3830 - val_accuracy: 0.8021\nEpoch 64/250\n561/561 - 0s - loss: 0.4835 - accuracy: 0.7861 - val_loss: 0.3840 - val_accuracy: 0.8021\nEpoch 65/250\n561/561 - 0s - loss: 0.4716 - accuracy: 0.7914 - val_loss: 0.3841 - val_accuracy: 0.8235\nEpoch 66/250\n561/561 - 0s - loss: 0.4787 - accuracy: 0.7861 - val_loss: 0.3831 - val_accuracy: 0.8182\nEpoch 67/250\n561/561 - 0s - loss: 0.4782 - accuracy: 0.7843 - val_loss: 0.3844 - val_accuracy: 0.8021\nEpoch 68/250\n561/561 - 0s - loss: 0.4766 - accuracy: 0.7968 - val_loss: 0.3867 - val_accuracy: 0.8289\nEpoch 69/250\n561/561 - 0s - loss: 0.4763 - accuracy: 0.7968 - val_loss: 0.3838 - val_accuracy: 0.8075\nEpoch 70/250\n561/561 - 0s - loss: 0.4748 - accuracy: 0.7825 - val_loss: 0.3848 - val_accuracy: 0.8182\nEpoch 71/250\n561/561 - 0s - loss: 0.4676 - accuracy: 0.7950 - val_loss: 0.3892 - val_accuracy: 0.8289\nEpoch 72/250\n561/561 - 0s - loss: 0.4802 - accuracy: 0.7932 - val_loss: 0.3843 - val_accuracy: 0.8075\nEpoch 73/250\n561/561 - 0s - loss: 0.4752 - accuracy: 0.7843 - val_loss: 0.3889 - val_accuracy: 0.7968\nEpoch 74/250\n561/561 - 0s - loss: 0.4802 - accuracy: 0.7754 - val_loss: 0.3862 - val_accuracy: 0.8235\nEpoch 75/250\n561/561 - 0s - loss: 0.4721 - accuracy: 0.8004 - val_loss: 0.3846 - val_accuracy: 0.8182\nEpoch 76/250\n561/561 - 0s - loss: 0.4762 - accuracy: 0.7914 - val_loss: 0.3835 - val_accuracy: 0.8075\nEpoch 77/250\n561/561 - 0s - loss: 0.4781 - accuracy: 0.7807 - val_loss: 0.3862 - val_accuracy: 0.8021\nEpoch 78/250\n561/561 - 0s - loss: 0.4727 - accuracy: 0.7932 - val_loss: 0.3858 - val_accuracy: 0.8128\nEpoch 79/250\n561/561 - 0s - loss: 0.4774 - accuracy: 0.7807 - val_loss: 0.3854 - val_accuracy: 0.8128\nEpoch 80/250\n561/561 - 0s - loss: 0.4747 - accuracy: 0.7932 - val_loss: 0.3860 - val_accuracy: 0.8021\nEpoch 81/250\n561/561 - 0s - loss: 0.4766 - accuracy: 0.7790 - val_loss: 0.3907 - val_accuracy: 0.8235\nEpoch 82/250\n561/561 - 0s - loss: 0.4761 - accuracy: 0.7950 - val_loss: 0.3857 - val_accuracy: 0.8128\nEpoch 83/250\n561/561 - 0s - loss: 0.4783 - accuracy: 0.7897 - val_loss: 0.3876 - val_accuracy: 0.8182\nEpoch 84/250\n561/561 - 0s - loss: 0.4735 - accuracy: 0.7897 - val_loss: 0.3866 - val_accuracy: 0.8021\nEpoch 85/250\n561/561 - 0s - loss: 0.4685 - accuracy: 0.7897 - val_loss: 0.3879 - val_accuracy: 0.8182\nEpoch 86/250\n561/561 - 0s - loss: 0.4855 - accuracy: 0.7986 - val_loss: 0.3855 - val_accuracy: 0.8128\nEpoch 87/250\n561/561 - 0s - loss: 0.4798 - accuracy: 0.7968 - val_loss: 0.3864 - val_accuracy: 0.8235\nEpoch 88/250\n561/561 - 0s - loss: 0.4832 - accuracy: 0.7986 - val_loss: 0.3857 - val_accuracy: 0.8128\nEpoch 89/250\n561/561 - 0s - loss: 0.4723 - accuracy: 0.7968 - val_loss: 0.3845 - val_accuracy: 0.8128\nEpoch 90/250\n561/561 - 0s - loss: 0.4830 - accuracy: 0.7807 - val_loss: 0.3844 - val_accuracy: 0.8021\nEpoch 91/250\n561/561 - 0s - loss: 0.4725 - accuracy: 0.7897 - val_loss: 0.3864 - val_accuracy: 0.8021\nEpoch 92/250\n561/561 - 0s - loss: 0.4843 - accuracy: 0.7950 - val_loss: 0.3874 - val_accuracy: 0.8128\nEpoch 93/250\n561/561 - 0s - loss: 0.4801 - accuracy: 0.7843 - val_loss: 0.3902 - val_accuracy: 0.8128\nEpoch 94/250\n561/561 - 0s - loss: 0.4856 - accuracy: 0.8004 - val_loss: 0.3862 - val_accuracy: 0.8021\nEpoch 95/250\n561/561 - 0s - loss: 0.4850 - accuracy: 0.7914 - val_loss: 0.3873 - val_accuracy: 0.8128\nEpoch 96/250\n561/561 - 0s - loss: 0.4720 - accuracy: 0.7914 - val_loss: 0.3864 - val_accuracy: 0.8021\nEpoch 97/250\n561/561 - 0s - loss: 0.4786 - accuracy: 0.7950 - val_loss: 0.3867 - val_accuracy: 0.7968\nEpoch 98/250\n561/561 - 0s - loss: 0.4828 - accuracy: 0.7790 - val_loss: 0.3889 - val_accuracy: 0.8235\nEpoch 99/250\n561/561 - 0s - loss: 0.4728 - accuracy: 0.7950 - val_loss: 0.3878 - val_accuracy: 0.8128\nEpoch 100/250\n561/561 - 0s - loss: 0.4820 - accuracy: 0.7932 - val_loss: 0.3882 - val_accuracy: 0.8182\nEpoch 101/250\n561/561 - 0s - loss: 0.4750 - accuracy: 0.7790 - val_loss: 0.3881 - val_accuracy: 0.7968\nEpoch 102/250\n561/561 - 0s - loss: 0.4733 - accuracy: 0.7879 - val_loss: 0.3879 - val_accuracy: 0.8021\nEpoch 103/250\n561/561 - 0s - loss: 0.4664 - accuracy: 0.7932 - val_loss: 0.3871 - val_accuracy: 0.8021\nEpoch 104/250\n561/561 - 0s - loss: 0.4700 - accuracy: 0.7932 - val_loss: 0.3880 - val_accuracy: 0.8128\nEpoch 105/250\n561/561 - 0s - loss: 0.4704 - accuracy: 0.7932 - val_loss: 0.3903 - val_accuracy: 0.8128\nEpoch 106/250\n561/561 - 0s - loss: 0.4705 - accuracy: 0.7879 - val_loss: 0.3878 - val_accuracy: 0.8021\nEpoch 107/250\n561/561 - 0s - loss: 0.4759 - accuracy: 0.7861 - val_loss: 0.3878 - val_accuracy: 0.8021\nEpoch 108/250\n561/561 - 0s - loss: 0.4806 - accuracy: 0.7897 - val_loss: 0.3880 - val_accuracy: 0.8021\nEpoch 109/250\n561/561 - 0s - loss: 0.4721 - accuracy: 0.7950 - val_loss: 0.3884 - val_accuracy: 0.8128\nEpoch 110/250\n561/561 - 0s - loss: 0.4733 - accuracy: 0.7932 - val_loss: 0.3892 - val_accuracy: 0.8128\nEpoch 111/250\n561/561 - 0s - loss: 0.4732 - accuracy: 0.7932 - val_loss: 0.3894 - val_accuracy: 0.8128\nEpoch 112/250\n561/561 - 0s - loss: 0.4716 - accuracy: 0.7968 - val_loss: 0.3899 - val_accuracy: 0.8128\nEpoch 113/250\n561/561 - 0s - loss: 0.4751 - accuracy: 0.7932 - val_loss: 0.3894 - val_accuracy: 0.8021\nEpoch 114/250\n561/561 - 0s - loss: 0.4752 - accuracy: 0.7897 - val_loss: 0.3900 - val_accuracy: 0.8128\nEpoch 115/250\n561/561 - 0s - loss: 0.4762 - accuracy: 0.7879 - val_loss: 0.3908 - val_accuracy: 0.8128\nEpoch 116/250\n561/561 - 0s - loss: 0.4846 - accuracy: 0.7807 - val_loss: 0.3913 - val_accuracy: 0.8128\nEpoch 117/250\n561/561 - 0s - loss: 0.4638 - accuracy: 0.7950 - val_loss: 0.3906 - val_accuracy: 0.8128\nEpoch 118/250\n561/561 - 0s - loss: 0.4612 - accuracy: 0.7968 - val_loss: 0.3894 - val_accuracy: 0.8128\nEpoch 119/250\n561/561 - 0s - loss: 0.4631 - accuracy: 0.7950 - val_loss: 0.3898 - val_accuracy: 0.8128\nEpoch 120/250\n561/561 - 0s - loss: 0.4657 - accuracy: 0.7932 - val_loss: 0.3890 - val_accuracy: 0.8128\nEpoch 121/250\n561/561 - 0s - loss: 0.4696 - accuracy: 0.8004 - val_loss: 0.3906 - val_accuracy: 0.8128\nEpoch 122/250\n561/561 - 0s - loss: 0.4686 - accuracy: 0.7932 - val_loss: 0.3902 - val_accuracy: 0.8128\nEpoch 123/250\n561/561 - 0s - loss: 0.4687 - accuracy: 0.7879 - val_loss: 0.3900 - val_accuracy: 0.8128\nEpoch 124/250\n561/561 - 0s - loss: 0.4727 - accuracy: 0.7843 - val_loss: 0.3912 - val_accuracy: 0.8128\nEpoch 125/250\n561/561 - 0s - loss: 0.4727 - accuracy: 0.7968 - val_loss: 0.3915 - val_accuracy: 0.8128\nEpoch 126/250\n561/561 - 0s - loss: 0.4723 - accuracy: 0.7932 - val_loss: 0.3908 - val_accuracy: 0.8128\nEpoch 127/250\n561/561 - 0s - loss: 0.4712 - accuracy: 0.7950 - val_loss: 0.3906 - val_accuracy: 0.8021\nEpoch 128/250\n561/561 - 0s - loss: 0.4762 - accuracy: 0.7914 - val_loss: 0.3900 - val_accuracy: 0.8021\nEpoch 129/250\n561/561 - 0s - loss: 0.4750 - accuracy: 0.7879 - val_loss: 0.3902 - val_accuracy: 0.8128\nEpoch 130/250\n561/561 - 0s - loss: 0.4733 - accuracy: 0.7861 - val_loss: 0.3907 - val_accuracy: 0.8128\nEpoch 131/250\n561/561 - 0s - loss: 0.4722 - accuracy: 0.7843 - val_loss: 0.3894 - val_accuracy: 0.7968\nEpoch 132/250\n561/561 - 0s - loss: 0.4802 - accuracy: 0.7897 - val_loss: 0.3915 - val_accuracy: 0.8128\nEpoch 133/250\n561/561 - 0s - loss: 0.4757 - accuracy: 0.7897 - val_loss: 0.3891 - val_accuracy: 0.7968\nEpoch 134/250\n561/561 - 0s - loss: 0.4768 - accuracy: 0.7879 - val_loss: 0.3898 - val_accuracy: 0.7968\nEpoch 135/250\n561/561 - 0s - loss: 0.4728 - accuracy: 0.7897 - val_loss: 0.3908 - val_accuracy: 0.8075\nEpoch 136/250\n561/561 - 0s - loss: 0.4700 - accuracy: 0.7914 - val_loss: 0.3893 - val_accuracy: 0.8021\nEpoch 137/250\n561/561 - 0s - loss: 0.4800 - accuracy: 0.7986 - val_loss: 0.3930 - val_accuracy: 0.8128\nEpoch 138/250\n561/561 - 0s - loss: 0.4621 - accuracy: 0.7968 - val_loss: 0.3919 - val_accuracy: 0.8128\nEpoch 139/250\n561/561 - 0s - loss: 0.4678 - accuracy: 0.7897 - val_loss: 0.3937 - val_accuracy: 0.8128\nEpoch 140/250\n561/561 - 0s - loss: 0.4740 - accuracy: 0.7790 - val_loss: 0.3933 - val_accuracy: 0.8075\nEpoch 141/250\n561/561 - 0s - loss: 0.4685 - accuracy: 0.8004 - val_loss: 0.3909 - val_accuracy: 0.8075\nEpoch 142/250\n561/561 - 0s - loss: 0.4716 - accuracy: 0.7986 - val_loss: 0.3913 - val_accuracy: 0.8128\nEpoch 143/250\n561/561 - 0s - loss: 0.4675 - accuracy: 0.7914 - val_loss: 0.3920 - val_accuracy: 0.8128\nEpoch 144/250\n561/561 - 0s - loss: 0.4726 - accuracy: 0.7932 - val_loss: 0.3921 - val_accuracy: 0.8128\nEpoch 145/250\n561/561 - 0s - loss: 0.4622 - accuracy: 0.8075 - val_loss: 0.3909 - val_accuracy: 0.8128\nEpoch 146/250\n561/561 - 0s - loss: 0.4777 - accuracy: 0.7968 - val_loss: 0.3896 - val_accuracy: 0.8128\nEpoch 147/250\n561/561 - 0s - loss: 0.4769 - accuracy: 0.7968 - val_loss: 0.3901 - val_accuracy: 0.8128\nEpoch 148/250\n561/561 - 0s - loss: 0.4705 - accuracy: 0.7914 - val_loss: 0.3910 - val_accuracy: 0.8128\nEpoch 149/250\n561/561 - 0s - loss: 0.4741 - accuracy: 0.7986 - val_loss: 0.3909 - val_accuracy: 0.8128\nEpoch 150/250\n561/561 - 0s - loss: 0.4722 - accuracy: 0.7772 - val_loss: 0.3925 - val_accuracy: 0.8128\nEpoch 151/250\n561/561 - 0s - loss: 0.4746 - accuracy: 0.7879 - val_loss: 0.3881 - val_accuracy: 0.8021\nEpoch 152/250\n561/561 - 0s - loss: 0.4816 - accuracy: 0.7932 - val_loss: 0.3900 - val_accuracy: 0.8021\nEpoch 153/250\n561/561 - 0s - loss: 0.4725 - accuracy: 0.7897 - val_loss: 0.3898 - val_accuracy: 0.8021\nEpoch 154/250\n561/561 - 0s - loss: 0.4747 - accuracy: 0.7986 - val_loss: 0.3910 - val_accuracy: 0.8128\nEpoch 155/250\n561/561 - 0s - loss: 0.4621 - accuracy: 0.7932 - val_loss: 0.3926 - val_accuracy: 0.8128\nEpoch 156/250\n561/561 - 0s - loss: 0.4800 - accuracy: 0.7879 - val_loss: 0.3903 - val_accuracy: 0.8128\nEpoch 157/250\n561/561 - 0s - loss: 0.4757 - accuracy: 0.8004 - val_loss: 0.3932 - val_accuracy: 0.8128\nEpoch 158/250\n561/561 - 0s - loss: 0.4741 - accuracy: 0.8057 - val_loss: 0.3924 - val_accuracy: 0.8128\nEpoch 159/250\n561/561 - 0s - loss: 0.4704 - accuracy: 0.7897 - val_loss: 0.3915 - val_accuracy: 0.8128\nEpoch 160/250\n561/561 - 0s - loss: 0.4776 - accuracy: 0.7950 - val_loss: 0.3894 - val_accuracy: 0.8021\nEpoch 161/250\n561/561 - 0s - loss: 0.4624 - accuracy: 0.7914 - val_loss: 0.3905 - val_accuracy: 0.8021\nEpoch 162/250\n561/561 - 0s - loss: 0.4673 - accuracy: 0.7861 - val_loss: 0.3907 - val_accuracy: 0.8128\nEpoch 163/250\n561/561 - 0s - loss: 0.4749 - accuracy: 0.7897 - val_loss: 0.3925 - val_accuracy: 0.8128\nEpoch 164/250\n561/561 - 0s - loss: 0.4707 - accuracy: 0.7932 - val_loss: 0.3941 - val_accuracy: 0.8128\nEpoch 165/250\n561/561 - 0s - loss: 0.4683 - accuracy: 0.7897 - val_loss: 0.3961 - val_accuracy: 0.8235\nEpoch 166/250\n561/561 - 0s - loss: 0.4738 - accuracy: 0.7950 - val_loss: 0.3927 - val_accuracy: 0.8128\nEpoch 167/250\n561/561 - 0s - loss: 0.4727 - accuracy: 0.7914 - val_loss: 0.3916 - val_accuracy: 0.8128\nEpoch 168/250\n561/561 - 0s - loss: 0.4637 - accuracy: 0.7861 - val_loss: 0.3940 - val_accuracy: 0.8128\nEpoch 169/250\n561/561 - 0s - loss: 0.4677 - accuracy: 0.7914 - val_loss: 0.3945 - val_accuracy: 0.8128\nEpoch 170/250\n561/561 - 0s - loss: 0.4705 - accuracy: 0.7968 - val_loss: 0.3921 - val_accuracy: 0.7968\nEpoch 171/250\n561/561 - 0s - loss: 0.4725 - accuracy: 0.7843 - val_loss: 0.3901 - val_accuracy: 0.8128\nEpoch 172/250\n561/561 - 0s - loss: 0.4701 - accuracy: 0.7950 - val_loss: 0.3927 - val_accuracy: 0.8128\nEpoch 173/250\n561/561 - 0s - loss: 0.4795 - accuracy: 0.7932 - val_loss: 0.3921 - val_accuracy: 0.8128\nEpoch 174/250\n561/561 - 0s - loss: 0.4742 - accuracy: 0.7879 - val_loss: 0.3896 - val_accuracy: 0.8128\nEpoch 175/250\n561/561 - 0s - loss: 0.4730 - accuracy: 0.7932 - val_loss: 0.3899 - val_accuracy: 0.8128\nEpoch 176/250\n561/561 - 0s - loss: 0.4713 - accuracy: 0.7914 - val_loss: 0.3930 - val_accuracy: 0.8128\nEpoch 177/250\n561/561 - 0s - loss: 0.4702 - accuracy: 0.7879 - val_loss: 0.3912 - val_accuracy: 0.8021\nEpoch 178/250\n561/561 - 0s - loss: 0.4687 - accuracy: 0.7772 - val_loss: 0.3905 - val_accuracy: 0.8021\nEpoch 179/250\n561/561 - 0s - loss: 0.4717 - accuracy: 0.7950 - val_loss: 0.3908 - val_accuracy: 0.8021\nEpoch 180/250\n561/561 - 0s - loss: 0.4730 - accuracy: 0.7879 - val_loss: 0.3922 - val_accuracy: 0.8128\nEpoch 181/250\n561/561 - 0s - loss: 0.4636 - accuracy: 0.7897 - val_loss: 0.3916 - val_accuracy: 0.7968\nEpoch 182/250\n561/561 - 0s - loss: 0.4747 - accuracy: 0.7772 - val_loss: 0.3928 - val_accuracy: 0.8128\nEpoch 183/250\n561/561 - 0s - loss: 0.4664 - accuracy: 0.7861 - val_loss: 0.3914 - val_accuracy: 0.7968\nEpoch 184/250\n561/561 - 0s - loss: 0.4720 - accuracy: 0.7950 - val_loss: 0.3916 - val_accuracy: 0.8128\nEpoch 185/250\n561/561 - 0s - loss: 0.4813 - accuracy: 0.7879 - val_loss: 0.3932 - val_accuracy: 0.8021\nEpoch 186/250\n561/561 - 0s - loss: 0.4750 - accuracy: 0.7914 - val_loss: 0.3921 - val_accuracy: 0.8128\nEpoch 187/250\n561/561 - 0s - loss: 0.4702 - accuracy: 0.7754 - val_loss: 0.3887 - val_accuracy: 0.8021\nEpoch 188/250\n561/561 - 0s - loss: 0.4668 - accuracy: 0.7932 - val_loss: 0.3902 - val_accuracy: 0.8128\nEpoch 189/250\n561/561 - 0s - loss: 0.4689 - accuracy: 0.7932 - val_loss: 0.3911 - val_accuracy: 0.8021\nEpoch 190/250\n561/561 - 0s - loss: 0.4660 - accuracy: 0.8021 - val_loss: 0.3915 - val_accuracy: 0.8128\nEpoch 191/250\n561/561 - 0s - loss: 0.4711 - accuracy: 0.7950 - val_loss: 0.3932 - val_accuracy: 0.8128\nEpoch 192/250\n561/561 - 0s - loss: 0.4766 - accuracy: 0.7861 - val_loss: 0.3935 - val_accuracy: 0.8021\nEpoch 193/250\n561/561 - 0s - loss: 0.4756 - accuracy: 0.7914 - val_loss: 0.3942 - val_accuracy: 0.8128\nEpoch 194/250\n561/561 - 0s - loss: 0.4742 - accuracy: 0.8004 - val_loss: 0.3922 - val_accuracy: 0.8021\nEpoch 195/250\n561/561 - 0s - loss: 0.4779 - accuracy: 0.7932 - val_loss: 0.3914 - val_accuracy: 0.7968\nEpoch 196/250\n561/561 - 0s - loss: 0.4668 - accuracy: 0.7932 - val_loss: 0.3937 - val_accuracy: 0.8128\nEpoch 197/250\n561/561 - 0s - loss: 0.4668 - accuracy: 0.7914 - val_loss: 0.3962 - val_accuracy: 0.8128\nEpoch 198/250\n561/561 - 0s - loss: 0.4660 - accuracy: 0.8004 - val_loss: 0.3928 - val_accuracy: 0.8021\nEpoch 199/250\n561/561 - 0s - loss: 0.4763 - accuracy: 0.7914 - val_loss: 0.3911 - val_accuracy: 0.8128\nEpoch 200/250\n561/561 - 0s - loss: 0.4692 - accuracy: 0.7968 - val_loss: 0.3931 - val_accuracy: 0.8128\nEpoch 201/250\n561/561 - 0s - loss: 0.4696 - accuracy: 0.7968 - val_loss: 0.3919 - val_accuracy: 0.8128\nEpoch 202/250\n561/561 - 0s - loss: 0.4745 - accuracy: 0.7807 - val_loss: 0.3920 - val_accuracy: 0.8128\nEpoch 203/250\n561/561 - 0s - loss: 0.4762 - accuracy: 0.7914 - val_loss: 0.3934 - val_accuracy: 0.7968\nEpoch 204/250\n561/561 - 0s - loss: 0.4684 - accuracy: 0.7897 - val_loss: 0.3945 - val_accuracy: 0.8128\nEpoch 205/250\n561/561 - 0s - loss: 0.4689 - accuracy: 0.7897 - val_loss: 0.3942 - val_accuracy: 0.8128\nEpoch 206/250\n561/561 - 0s - loss: 0.4704 - accuracy: 0.7968 - val_loss: 0.3927 - val_accuracy: 0.8128\nEpoch 207/250\n561/561 - 0s - loss: 0.4696 - accuracy: 0.7861 - val_loss: 0.3946 - val_accuracy: 0.8128\nEpoch 208/250\n561/561 - 0s - loss: 0.4653 - accuracy: 0.7986 - val_loss: 0.3957 - val_accuracy: 0.8128\nEpoch 209/250\n561/561 - 0s - loss: 0.4765 - accuracy: 0.8004 - val_loss: 0.3954 - val_accuracy: 0.8128\nEpoch 210/250\n561/561 - 0s - loss: 0.4707 - accuracy: 0.7861 - val_loss: 0.3931 - val_accuracy: 0.8021\nEpoch 211/250\n561/561 - 0s - loss: 0.4711 - accuracy: 0.7897 - val_loss: 0.3932 - val_accuracy: 0.8128\nEpoch 212/250\n561/561 - 0s - loss: 0.4613 - accuracy: 0.7986 - val_loss: 0.3935 - val_accuracy: 0.8128\nEpoch 213/250\n561/561 - 0s - loss: 0.4677 - accuracy: 0.7914 - val_loss: 0.3936 - val_accuracy: 0.8128\nEpoch 214/250\n561/561 - 0s - loss: 0.4788 - accuracy: 0.7986 - val_loss: 0.3917 - val_accuracy: 0.8128\nEpoch 215/250\n561/561 - 0s - loss: 0.4812 - accuracy: 0.7914 - val_loss: 0.3920 - val_accuracy: 0.8128\nEpoch 216/250\n561/561 - 0s - loss: 0.4766 - accuracy: 0.7879 - val_loss: 0.3944 - val_accuracy: 0.7968\nEpoch 217/250\n561/561 - 0s - loss: 0.4724 - accuracy: 0.7861 - val_loss: 0.3942 - val_accuracy: 0.8128\nEpoch 218/250\n561/561 - 0s - loss: 0.4626 - accuracy: 0.7932 - val_loss: 0.3944 - val_accuracy: 0.8128\nEpoch 219/250\n561/561 - 0s - loss: 0.4722 - accuracy: 0.7932 - val_loss: 0.3939 - val_accuracy: 0.8128\nEpoch 220/250\n561/561 - 0s - loss: 0.4775 - accuracy: 0.7843 - val_loss: 0.3917 - val_accuracy: 0.7968\nEpoch 221/250\n561/561 - 0s - loss: 0.4768 - accuracy: 0.7790 - val_loss: 0.3940 - val_accuracy: 0.8128\nEpoch 222/250\n561/561 - 0s - loss: 0.4723 - accuracy: 0.7914 - val_loss: 0.3917 - val_accuracy: 0.8021\nEpoch 223/250\n561/561 - 0s - loss: 0.4752 - accuracy: 0.7914 - val_loss: 0.3927 - val_accuracy: 0.8128\nEpoch 224/250\n561/561 - 0s - loss: 0.4709 - accuracy: 0.7932 - val_loss: 0.3941 - val_accuracy: 0.8128\nEpoch 225/250\n561/561 - 0s - loss: 0.4696 - accuracy: 0.7879 - val_loss: 0.3945 - val_accuracy: 0.8021\nEpoch 226/250\n561/561 - 0s - loss: 0.4738 - accuracy: 0.7772 - val_loss: 0.3908 - val_accuracy: 0.8128\nEpoch 227/250\n561/561 - 0s - loss: 0.4762 - accuracy: 0.7914 - val_loss: 0.3928 - val_accuracy: 0.8128\nEpoch 228/250\n561/561 - 0s - loss: 0.4693 - accuracy: 0.7968 - val_loss: 0.3954 - val_accuracy: 0.8128\nEpoch 229/250\n561/561 - 0s - loss: 0.4643 - accuracy: 0.7879 - val_loss: 0.3921 - val_accuracy: 0.8021\nEpoch 230/250\n561/561 - 0s - loss: 0.4679 - accuracy: 0.7879 - val_loss: 0.3933 - val_accuracy: 0.8128\nEpoch 231/250\n561/561 - 0s - loss: 0.4702 - accuracy: 0.7968 - val_loss: 0.3922 - val_accuracy: 0.8128\nEpoch 232/250\n561/561 - 0s - loss: 0.4795 - accuracy: 0.7790 - val_loss: 0.3927 - val_accuracy: 0.8128\nEpoch 233/250\n561/561 - 0s - loss: 0.4630 - accuracy: 0.7968 - val_loss: 0.3911 - val_accuracy: 0.8128\nEpoch 234/250\n561/561 - 0s - loss: 0.4741 - accuracy: 0.7914 - val_loss: 0.3909 - val_accuracy: 0.8128\nEpoch 235/250\n561/561 - 0s - loss: 0.4708 - accuracy: 0.7897 - val_loss: 0.3951 - val_accuracy: 0.8128\nEpoch 236/250\n561/561 - 0s - loss: 0.4698 - accuracy: 0.7843 - val_loss: 0.3917 - val_accuracy: 0.8021\nEpoch 237/250\n561/561 - 0s - loss: 0.4639 - accuracy: 0.8004 - val_loss: 0.3915 - val_accuracy: 0.8128\nEpoch 238/250\n561/561 - 0s - loss: 0.4624 - accuracy: 0.7897 - val_loss: 0.3926 - val_accuracy: 0.8128\nEpoch 239/250\n561/561 - 0s - loss: 0.4736 - accuracy: 0.7968 - val_loss: 0.3930 - val_accuracy: 0.8128\nEpoch 240/250\n561/561 - 0s - loss: 0.4684 - accuracy: 0.7950 - val_loss: 0.3929 - val_accuracy: 0.8128\nEpoch 241/250\n561/561 - 0s - loss: 0.4701 - accuracy: 0.7807 - val_loss: 0.3938 - val_accuracy: 0.8128\nEpoch 242/250\n561/561 - 0s - loss: 0.4694 - accuracy: 0.7950 - val_loss: 0.3945 - val_accuracy: 0.8128\nEpoch 243/250\n561/561 - 0s - loss: 0.4629 - accuracy: 0.7968 - val_loss: 0.3938 - val_accuracy: 0.8128\nEpoch 244/250\n561/561 - 0s - loss: 0.4716 - accuracy: 0.7861 - val_loss: 0.3961 - val_accuracy: 0.8128\nEpoch 245/250\n561/561 - 0s - loss: 0.4740 - accuracy: 0.7932 - val_loss: 0.3940 - val_accuracy: 0.8128\nEpoch 246/250\n561/561 - 0s - loss: 0.4558 - accuracy: 0.8004 - val_loss: 0.3961 - val_accuracy: 0.8128\nEpoch 247/250\n561/561 - 0s - loss: 0.4652 - accuracy: 0.7861 - val_loss: 0.3941 - val_accuracy: 0.8128\nEpoch 248/250\n561/561 - 0s - loss: 0.4741 - accuracy: 0.7807 - val_loss: 0.3915 - val_accuracy: 0.8021\nEpoch 249/250\n561/561 - 0s - loss: 0.4799 - accuracy: 0.7861 - val_loss: 0.3945 - val_accuracy: 0.8128\nEpoch 250/250\n561/561 - 0s - loss: 0.4723 - accuracy: 0.7968 - val_loss: 0.3948 - val_accuracy: 0.8021\n" ], [ "", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Now You Code 2: Is That An Email Address?\n\nLet's use Python's built-in string functions to write our own function to detect if a string is an email address. \n\nThe function `isEmail(text)` should return `True` when `text` is an email address, `False` otherwise. \n\nFor simplicity's sake we will define an email address to be any string with just ONE `@` symbol in it, where the `@` is not at the beginning or end of the string. So `a@b` is considered an email (even though it really isn't).\n\nThe program should detect emails until you enter quit. \n\nSample run:\n```\nEmail address detector. Type quit to exit. \nEmail: [email protected]\[email protected] ==> email\nEmail: mafudge@\nmafudge@ ==> NOT EMAIL\nEmail: mafudge\nmafudge ==> NOT EMAIL\nEmail: @syr.edu\[email protected] ==> NOT EMAIL\nEmail: @\n@ ==> NOT EMAIL\nEmail: mafudge@@syr.edu\nmafudge@@syr.edu ==> NOT EMAIL\nEmail: mafudge@syr@edu\nmafudge@syr@edu ==> NOT EMAIL\n```\n\nOnce again we will use the problem simplification technique to writing this program.\n\nFirst we will write the `isEmail(text)` function, then we will write the main program.\n", "_____no_output_____" ], [ "## Step 1: Problem Analysis for isEmail function only\n\nInputs (function arguments):\n\nOutputs (what is returns): \n\nAlgorithm (Steps in Function):\n\n", "_____no_output_____" ] ], [ [ "## Step 2: Todo write the function definition for isEmail functiuon\n", "_____no_output_____" ], [ "## Step 3: Write some tests, to ensure the function works, for example\n## Make sure to test all cases!\nprint(\"WHEN [email protected] We EXPECT isEmail(text) to return True\", \"ACTUAL\", isEmail(\"[email protected]\") )\nprint(\"WHEN text=mike@ We EXPECT isEmail(text) to return False\", \"ACTUAL\", isEmail(\"mike@\") )\n\n", "_____no_output_____" ] ], [ [ "## Step 4: Problem Analysis for full Program\n\nInputs:\n\nOutputs:\n\nAlgorithm (Steps in Program):\n\n", "_____no_output_____" ] ], [ [ "## Step 5: todo write code for full problem, using the isEmail function to help you solve the problem\n", "_____no_output_____" ] ], [ [ "## Step 6: Questions\n\n1. How many test cases should you have in step 3 to ensure you've tested all the cases?\n2. What should kind of logic should we add to make our `isEmail` function even better, so that is detects emails more accurately?\n", "_____no_output_____" ], [ "## Reminder of Evaluation Criteria\n\n1. What the problem attempted (analysis, code, and answered questions) ?\n2. What the problem analysis thought out? (does the program match the plan?)\n3. Does the code execute without syntax error?\n4. Does the code solve the intended problem?\n5. Is the code well written? (easy to understand, modular, and self-documenting, handles errors)\n", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "import numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\nimport seaborn as sns\n%matplotlib inline", "_____no_output_____" ], [ "from sklearn.datasets import load_breast_cancer", "_____no_output_____" ], [ "cancer = load_breast_cancer()", "_____no_output_____" ], [ "cancer.keys()", "_____no_output_____" ], [ "print(cancer['DESCR'])", ".. _breast_cancer_dataset:\n\nBreast cancer wisconsin (diagnostic) dataset\n--------------------------------------------\n\n**Data Set Characteristics:**\n\n :Number of Instances: 569\n\n :Number of Attributes: 30 numeric, predictive attributes and the class\n\n :Attribute Information:\n - radius (mean of distances from center to points on the perimeter)\n - texture (standard deviation of gray-scale values)\n - perimeter\n - area\n - smoothness (local variation in radius lengths)\n - compactness (perimeter^2 / area - 1.0)\n - concavity (severity of concave portions of the contour)\n - concave points (number of concave portions of the contour)\n - symmetry \n - fractal dimension (\"coastline approximation\" - 1)\n\n The mean, standard error, and \"worst\" or largest (mean of the three\n largest values) of these features were computed for each image,\n resulting in 30 features. For instance, field 3 is Mean Radius, field\n 13 is Radius SE, field 23 is Worst Radius.\n\n - class:\n - WDBC-Malignant\n - WDBC-Benign\n\n :Summary Statistics:\n\n ===================================== ====== ======\n Min Max\n ===================================== ====== ======\n radius (mean): 6.981 28.11\n texture (mean): 9.71 39.28\n perimeter (mean): 43.79 188.5\n area (mean): 143.5 2501.0\n smoothness (mean): 0.053 0.163\n compactness (mean): 0.019 0.345\n concavity (mean): 0.0 0.427\n concave points (mean): 0.0 0.201\n symmetry (mean): 0.106 0.304\n fractal dimension (mean): 0.05 0.097\n radius (standard error): 0.112 2.873\n texture (standard error): 0.36 4.885\n perimeter (standard error): 0.757 21.98\n area (standard error): 6.802 542.2\n smoothness (standard error): 0.002 0.031\n compactness (standard error): 0.002 0.135\n concavity (standard error): 0.0 0.396\n concave points (standard error): 0.0 0.053\n symmetry (standard error): 0.008 0.079\n fractal dimension (standard error): 0.001 0.03\n radius (worst): 7.93 36.04\n texture (worst): 12.02 49.54\n perimeter (worst): 50.41 251.2\n area (worst): 185.2 4254.0\n smoothness (worst): 0.071 0.223\n compactness (worst): 0.027 1.058\n concavity (worst): 0.0 1.252\n concave points (worst): 0.0 0.291\n symmetry (worst): 0.156 0.664\n fractal dimension (worst): 0.055 0.208\n ===================================== ====== ======\n\n :Missing Attribute Values: None\n\n :Class Distribution: 212 - Malignant, 357 - Benign\n\n :Creator: Dr. William H. Wolberg, W. Nick Street, Olvi L. Mangasarian\n\n :Donor: Nick Street\n\n :Date: November, 1995\n\nThis is a copy of UCI ML Breast Cancer Wisconsin (Diagnostic) datasets.\nhttps://goo.gl/U2Uwz2\n\nFeatures are computed from a digitized image of a fine needle\naspirate (FNA) of a breast mass. They describe\ncharacteristics of the cell nuclei present in the image.\n\nSeparating plane described above was obtained using\nMultisurface Method-Tree (MSM-T) [K. P. Bennett, \"Decision Tree\nConstruction Via Linear Programming.\" Proceedings of the 4th\nMidwest Artificial Intelligence and Cognitive Science Society,\npp. 97-101, 1992], a classification method which uses linear\nprogramming to construct a decision tree. Relevant features\nwere selected using an exhaustive search in the space of 1-4\nfeatures and 1-3 separating planes.\n\nThe actual linear program used to obtain the separating plane\nin the 3-dimensional space is that described in:\n[K. P. Bennett and O. L. Mangasarian: \"Robust Linear\nProgramming Discrimination of Two Linearly Inseparable Sets\",\nOptimization Methods and Software 1, 1992, 23-34].\n\nThis database is also available through the UW CS ftp server:\n\nftp ftp.cs.wisc.edu\ncd math-prog/cpo-dataset/machine-learn/WDBC/\n\n.. topic:: References\n\n - W.N. Street, W.H. Wolberg and O.L. Mangasarian. Nuclear feature extraction \n for breast tumor diagnosis. IS&T/SPIE 1993 International Symposium on \n Electronic Imaging: Science and Technology, volume 1905, pages 861-870,\n San Jose, CA, 1993.\n - O.L. Mangasarian, W.N. Street and W.H. Wolberg. Breast cancer diagnosis and \n prognosis via linear programming. Operations Research, 43(4), pages 570-577, \n July-August 1995.\n - W.H. Wolberg, W.N. Street, and O.L. Mangasarian. Machine learning techniques\n to diagnose breast cancer from fine-needle aspirates. Cancer Letters 77 (1994) \n 163-171.\n" ], [ "cancer['feature_names']", "_____no_output_____" ], [ "df_feat = pd.DataFrame(cancer['data'],columns=cancer['feature_names'])\ndf_feat.info()", "<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 569 entries, 0 to 568\nData columns (total 30 columns):\nmean radius 569 non-null float64\nmean texture 569 non-null float64\nmean perimeter 569 non-null float64\nmean area 569 non-null float64\nmean smoothness 569 non-null float64\nmean compactness 569 non-null float64\nmean concavity 569 non-null float64\nmean concave points 569 non-null float64\nmean symmetry 569 non-null float64\nmean fractal dimension 569 non-null float64\nradius error 569 non-null float64\ntexture error 569 non-null float64\nperimeter error 569 non-null float64\narea error 569 non-null float64\nsmoothness error 569 non-null float64\ncompactness error 569 non-null float64\nconcavity error 569 non-null float64\nconcave points error 569 non-null float64\nsymmetry error 569 non-null float64\nfractal dimension error 569 non-null float64\nworst radius 569 non-null float64\nworst texture 569 non-null float64\nworst perimeter 569 non-null float64\nworst area 569 non-null float64\nworst smoothness 569 non-null float64\nworst compactness 569 non-null float64\nworst concavity 569 non-null float64\nworst concave points 569 non-null float64\nworst symmetry 569 non-null float64\nworst fractal dimension 569 non-null float64\ndtypes: float64(30)\nmemory usage: 133.4 KB\n" ], [ "cancer['target']", "_____no_output_____" ], [ "df_target = pd.DataFrame(cancer['target'],columns=['Cancer'])", "_____no_output_____" ], [ "df_feat.head()", "_____no_output_____" ], [ "from sklearn.model_selection import train_test_split", "_____no_output_____" ], [ "X_train, X_test, y_train, y_test = train_test_split(df_feat, np.ravel(df_target), test_size=0.30, random_state=101)", "_____no_output_____" ], [ "from sklearn.svm import SVC\nmodel = SVC()\nmodel.fit(X_train,y_train)", "C:\\Users\\KIIT\\Anaconda3\\lib\\site-packages\\sklearn\\svm\\base.py:196: FutureWarning: The default value of gamma will change from 'auto' to 'scale' in version 0.22 to account better for unscaled features. Set gamma explicitly to 'auto' or 'scale' to avoid this warning.\n \"avoid this warning.\", FutureWarning)\n" ], [ "predictions = model.predict(X_test)", "_____no_output_____" ], [ "from sklearn.metrics import classification_report,confusion_matrix", "_____no_output_____" ], [ "print(confusion_matrix(y_test,predictions))", "[[ 0 66]\n [ 0 105]]\n" ], [ "print(classification_report(y_test,predictions))", " precision recall f1-score support\n\n 0 0.00 0.00 0.00 66\n 1 0.61 1.00 0.76 105\n\n micro avg 0.61 0.61 0.61 171\n macro avg 0.31 0.50 0.38 171\nweighted avg 0.38 0.61 0.47 171\n\n" ], [ "param_grid = {'C': [0.1,1, 10, 100, 1000], 'gamma': [1,0.1,0.01,0.001,0.0001], 'kernel': ['rbf']}", "_____no_output_____" ], [ "from sklearn.model_selection import GridSearchCV\ngrid = GridSearchCV(SVC(),param_grid,refit=True,verbose=3)\ngrid.fit(X_train,y_train)", "C:\\Users\\KIIT\\Anaconda3\\lib\\site-packages\\sklearn\\model_selection\\_split.py:2053: FutureWarning: You should specify a value for 'cv' instead of relying on the default value. The default value will change from 3 to 5 in version 0.22.\n warnings.warn(CV_WARNING, FutureWarning)\n[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n[Parallel(n_jobs=1)]: Done 1 out of 1 | elapsed: 0.0s remaining: 0.0s\n[Parallel(n_jobs=1)]: Done 2 out of 2 | elapsed: 0.0s remaining: 0.0s\n" ], [ "grid.best_params_", "_____no_output_____" ], [ "grid.best_estimator_", "_____no_output_____" ], [ "grid_predictions = grid.predict(X_test)\nprint(confusion_matrix(y_test,grid_predictions))\nprint(classification_report(y_test,grid_predictions))", "[[ 60 6]\n [ 3 102]]\n precision recall f1-score support\n\n 0 0.95 0.91 0.93 66\n 1 0.94 0.97 0.96 105\n\n micro avg 0.95 0.95 0.95 171\n macro avg 0.95 0.94 0.94 171\nweighted avg 0.95 0.95 0.95 171\n\n" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Apache 2.0 License\n# Copyright (c) 2022, Fraunhofer e.V.\n# All rights reserved.\n\nimport sys\nsys.path.append('../')\nimport argparse\nfrom tqdm import tqdm\nfrom scipy.stats import entropy\nimport torch\nimport torch.nn.functional as F\n\nfrom timm.models import create_model\nfrom defenses.victim import MAD, ReverseSigmoid, RandomNoise\n\nfrom datasets import build_transform, get_dataset\n\nimport models\nimport utils\nfrom utils import get_free_gpu\n\nnum_gpus = 1\ngpu_chosen = get_free_gpu(num_gpus)\ndevice = torch.device('cuda' if torch.cuda.is_available() else \"cpu\")", "_____no_output_____" ], [ "args = {'input_size': 224}\nargs = argparse.Namespace(**args)", "_____no_output_____" ], [ "def get_accuracy(output, target, topk=(1,)):\n \"\"\" Computes the precision@k for the specified values of k \"\"\"\n maxk = max(topk)\n batch_size = target.size(0)\n\n _, pred = output.topk(maxk, 1, True, True)\n pred = pred.t()\n # one-hot case\n if target.ndimension() > 1:\n target = target.max(1)[1]\n\n correct = pred.eq(target.view(1, -1).expand_as(pred))\n\n res = dict()\n for k in topk:\n correct_k = correct[:k].reshape(-1).float().sum(0)\n res[\"acc{}\".format(k)] = correct_k.mul_(1.0 / batch_size).item()\n return res\n\n\ndef predict(model, model_defended, data_loader, device='cuda'):\n model = model.to(device)\n model.eval()\n preds_orig = []\n preds_def = []\n labels = []\n with torch.no_grad():\n for x, y in tqdm(data_loader):\n x, y = x.to(device, non_blocking=True), y.to(device, non_blocking=True)\n preds_orig.append(F.softmax(model(x), 1).to('cpu'))\n preds_def.append(model_defended(x).to('cpu'))\n labels.append(y.to('cpu'))\n return torch.cat(preds_orig), torch.cat(preds_def), torch.cat(labels)\n\n\ndef evaluate(model, model_defended, datasets, batch_size=100, workers=4):\n if not isinstance(datasets, tuple):\n datasets = (datasets, )\n res = {}\n for i, dataset in enumerate(datasets):\n d_type = \"\" if len(datasets) == 1 else [\"train\", \"test\"][i]\n data_loader = torch.utils.data.DataLoader(dataset, batch_size=batch_size, num_workers=workers)\n \n print(f'Evaluate on {dataset.__class__.__name__} {d_type} data:')\n preds_orig, preds_def, labels = predict(model, model_defended, data_loader)\n num_classes = preds_def.shape[1]\n \n print(f'Results on {dataset.__class__.__name__} {d_type} data:')\n print('Accuracy original:', get_accuracy(preds_orig, labels)['acc1'])\n print('Accuracy defended:', get_accuracy(preds_def, labels)['acc1'])\n print('Fidelity:', get_accuracy(preds_orig, preds_def)['acc1'])\n print('Mean relative entropy original:', np.mean(entropy(preds_orig, axis=1, base=2) / np.log2(num_classes)))\n print('Mean relative entropy defended:', np.mean(entropy(preds_def, axis=1, base=2) / np.log2(num_classes)))\n print('Mean max/min original:', torch.mean(preds_orig.max(1)[0] / preds_orig.min(1)[0]).item())\n print('Mean max/min defended:', torch.mean(preds_def.max(1)[0] / preds_def.min(1)[0]).item())\n print('Mean L1 distance:', torch.mean(torch.linalg.vector_norm(preds_orig - preds_def, 1, 1)).item())\n print()\n res[d_type] = (preds_orig, preds_def, labels)\n return res", "_____no_output_____" ], [ "model = create_model(\n 'resnet34',\n num_classes=10\n)\nmodel.load_state_dict(torch.load(f'checkpoints/checkpoint.pth')['model'])\n\nmodel_adv = create_model(\n 'deit_base_patch16_224',\n pretrained=False,\n num_classes=10\n)\n\ndatasets = get_dataset('cifar10', train_transform=build_transform(False, args), val_transform=build_transform(False, args))", "_____no_output_____" ], [ "epsilon = 50.1\nbeta = 2.0\ngamma = 0.5\noutput_path = './log'\ndist_z = 'l1'\noracle = 'argmax'\n\nmodel_defended = RandomNoise(model=model, out_path=output_path, dist_z=dist_z, epsilon_z=epsilon)\n#model_defended = ReverseSigmoid(model=model, out_path=output_path, beta=beta, gamma=gamma)\n#model_defended = MAD(model=model, out_path=output_path, epsilon=epsilon, model_adv_proxy=model_adv, oracle=oracle)", "_____no_output_____" ], [ "res = evaluate(model, model_defended, datasets[1], 512)", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "from numpy import *\nfrom numpy.random import *\nimport matplotlib.pyplot as plt\nimport matplotlib.cm as cm\nfrom PlotFuncs import *\nfrom LabFuncs import *\nfrom Params import *\nfrom HaloFuncs import *\nfrom scipy.stats import norm\nimport pandas\n\nimport cmocean\n\n# Load shards\ndf = pandas.read_csv('../data/FitShards_red.csv')\nnames = df.group_id\nnshards = size(names)\nvelocities = zeros(shape=(nshards,3))\ndispersions = zeros(shape=(nshards,3))\nvelocities[0:(nshards),0] = df.vx # stream velocities\nvelocities[0:(nshards),1] = df.vy\nvelocities[0:(nshards),2] = df.vz\ndispersions[0:(nshards),0] = df.sigx # dispersion tensors\ndispersions[0:(nshards),1] = df.sigy\ndispersions[0:(nshards),2] = df.sigz\npops = df.population\nPsun = df.Psun\nweights = ShardsWeights(names,pops,Psun)\n\n# Halo params\nHaloModel = SHMpp\nv0 = HaloModel.RotationSpeed\nv_esc = HaloModel.EscapeSpeed\nsig_beta = HaloModel.SausageDispersionTensor\nsig_iso = array([1.0,1.0,1.0])*v0/sqrt(2.0)\nday = 67.0 ", "_____no_output_____" ], [ "from scipy.constants import c\nc", "_____no_output_____" ], [ "# Angular discretisation\n\n# n = 50\n# nn = int(n*n/2)\n# ph = linspace(0, 2*pi, n)\n# costh = linspace(-1.0, 1.0, int(n/2))\n# PH, cosTH = meshgrid(ph,costh)\n# X = sin(PH)*cosTH\n# Y = sin(PH)*sqrt(1-cosTH**2.0)\n# Z = cosTH\n# ii = 0\n# q = zeros(shape=(nn,3))\n# for i in range(0,int(n/2)):\n# for j in range(0,n):\n# q[ii,:] = array([X[i,j],Y[i,j],Z[i,j]])\n# ii = ii+1\n\nimport healpy as hp\n\nnside = 64\nnn = 12*nside**2\ndpix = 4*pi/(npix*1.0)\nq = zeros(shape=(nn,3))\nfor i in range(0,nn):\n q[i,:] = hp.pix2vec(nside, i)\n \nforward = q[:,1]<0\nbackward = ~forward\nup = q[:,2]>0\ndown = ~up\n", "_____no_output_____" ], [ "from WIMPFuncs import *\n\nvvlab = sqrt(sum(LabVelocitySimple(day)**2.0))\nvlab = array([0.0,vvlab,0.0])\nvshift = LabVelocitySimple(day)-vlab\nxlab = vlab/vvlab\n\n\n\n\n# constants\nsigma_p = 1.0e-46\nA = F19.MassNumber\neta = 0.2\n\n\n# Radon transforms\nnv = 500\nv_min = linspace(0.01,850.0,nv)\nfhat_smooth = zeros(shape=(nn,nv))\nfhat_shards = zeros(shape=(nn,nv))\nfor iv in range(0,nv):\n fhat_smooth[:,iv] = (1-eta)*fhat_Isotropic(v_min[iv],q,day,v_shift=vshift)\\\n +eta*fhat_Triaxial(v_min[iv],q,day,sig_beta,v_shift=vshift)\n# fhat = zeros(shape=(nn))\n# for i in range(0,nshards):\n# v_s = velocities[i,:]\n# sig_s = dispersions[i,:]\n# fhat += weights[i]*fhat_Triaxial(v_min[iv],q,day,sig_s,v_shift=v_s+vshift)\n# fhat_shards[:,iv] = fhat\n \n #fhat_shards[:,iv] = fhat_Triaxial(v_min[iv],q,day,dispersions[2,:],v_shift=velocities[2,:]+vshift)\n fhat_shards[:,iv] = fhat_Triaxial(v_min[iv],q,day,dispersions[0,:],v_shift=velocities[0,:]+vshift)\n \n\n\n", "_____no_output_____" ], [ "# RATEs\n\ndef CAnisotropies(m_chi,v_min,q,fhat,E_r):\n R = zeros(shape=shape(fhat)[0])\n ne = size(E_r)\n for i in range(0,ne):\n v_min_i = MinimumWIMPSpeed(E_r[i],F19.MassNumber,m_chi)\n R += diffRecoilRate_SI(E_r[i],fhat[:,argmin(abs(v_min-v_min_i))],F19.MassNumber,sigma_p,m_chi)\n R_fw = sum(R[forward])\n R_bw = sum(R[backward])\n R_up = sum(R[up])\n R_dn = sum(R[down])\n return R_fw,R_bw,R_up,R_dn\n\n \n \n# E range\nne = 100\nE_r = zeros(shape=ne)\nE_r = linspace(2.0,50.0,ne)\n\n# xi range\nn_xi = 10\nxi_vals = linspace(0.0,0.2,n_xi)\n\n# m range\nnm = 50\nm_vals = logspace(log10(1.0),log10(1000.0),nm)\n\n\nFW = zeros(shape=(nm,n_xi))\nBW = zeros(shape=(nm,n_xi))\nUP = zeros(shape=(nm,n_xi))\nDN = zeros(shape=(nm,n_xi))\n\nfor i in range(0,nm):\n R0_fw,R0_bw,R0_up,R0_dn = CAnisotropies(m_vals[i],v_min,q,fhat_smooth,E_r)\n R1_fw,R1_bw,R1_up,R1_dn = CAnisotropies(m_vals[i],v_min,q,fhat_shards,E_r)\n for j in range(0,n_xi):\n xi = xi_vals[j]\n FW[i,j] = R0_fw*(1-xi)+R1_fw*xi\n BW[i,j] = R0_bw*(1-xi)+R1_bw*xi\n UP[i,j] = R0_up*(1-xi)+R1_up*xi\n DN[i,j] = R0_dn*(1-xi)+R1_dn*xi", "_____no_output_____" ], [ " \n\nfig,ax1 = MySquarePlot('$m_\\chi$ [GeV]','Anisotropy',lfs=40,tfs=30)\ncol = cmocean.cm.curl(linspace(0,1,n_xi))\n\nfor i in flipud(range(0,n_xi)):\n plt.fill_between(m_vals,FW[:,i]/BW[:,i],y2=FW[:,0]/BW[:,0],lw=3,color=col[i,:])\n plt.fill_between(m_vals,DN[:,i]/UP[:,i],y2=DN[:,0]/UP[:,0],lw=3,color=col[i,:])\n #plt.plot(m_vals,FW[:,i]/BW[:,i],'-',lw=5,color='k')\n #plt.plot(m_vals,DN[:,i]/UP[:,i],'-',lw=5,color='k')\n plt.plot(m_vals,FW[:,i]/BW[:,i],'-',lw=3,color=col[i,:])\n plt.plot(m_vals,DN[:,i]/UP[:,i],'-',lw=3,color=col[i,:])\n \n \n\nplt.gcf().text(0.53,0.7,r'$\\mathcal{A}_{\\rm fw}$',fontsize=45,color=col[0,:])\nplt.gcf().text(0.4,0.14,r'$\\mathcal{A}_{\\rm ud}$',fontsize=45,color=col[0,:])\n\nymin = 1.0\nymax = 10.0\nxmin = 5.0\nxmax = 1000.0\n\n#ivals = array(11,)\n\nNvals = array([14,15,16,17,18,19,20,23,30,45,100,250])\nfor i in Nvals:\n A = (sqrt(i*1)/3.0 + 1)/(sqrt(i*1)/3.0 - 1)\n #plt.plot([xmin,xmax],[A,A],'--',lw=2,color='gray')\n plt.plot([xmax*0.95,xmax],[A,A],'-',lw=3,color='k')\n plt.text(xmax+20,A,str(i),fontsize=30,va='center')\n\nplt.ylim([ymin,ymax])\nplt.xscale('log')\nplt.xlim([xmin,xmax])\nplt.ylim([ymin,ymax])\n\n\n#ax2 = ax1.twinx()\n#ax2.set_yscale('log')\n#ax2.set_ylim([Niso(ymin),Niso(ymax)])\n#ax2.set_yticks(arange(Niso(ymin),Niso(ymax),1))\n\n#ax1.xaxis.grid() # vertical lines\n#ax1.xaxis.grid(which='minor') # vertical lines\n\nax1.tick_params(which='major',right=False)\nax1.tick_params(which='minor',right=False)\n\n\nplt.gcf().text(0.93,0.45,r'$N_{\\rm iso}$',fontsize=45,rotation=-90)\n\n\n# Custom colorbar\ncbar_max = r'10 \\% Shards'\ncbar_min = r'0\\% Shards'\nimport matplotlib as mpl\nfrom mpl_toolkits.axes_grid1.inset_locator import inset_axes\ncbaxes = inset_axes(ax1, width=\"15%\", height=\"30%\", bbox_to_anchor=[0, -330, 200, 760]) \nnorm = mpl.colors.Normalize(vmin=0,vmax=1)\nsm = plt.cm.ScalarMappable(cmap=cmocean.cm.curl, norm=norm)\nsm.set_array([])\nplt.colorbar(sm,cax=cbaxes,ticks=(0,1),boundaries=linspace(0,1,n_xi),orientation='vertical')\nf = plt.gcf().get_children()\ncbar = f[2]\ncbar.set_yticklabels([cbar_min,cbar_max]) # vertically oriented colorbar\ncbar.tick_params(labelsize=30) \ncbar.tick_params(which='major',direction='out',width=2,length=10,right=True,top=False,pad=7)\ncbar.tick_params(which='minor',direction='out',width=2,length=7,right=True,top=False)\nplt.gcf().text(0.23,0.33,r'$\\xi_{\\rm S1}$',fontsize=35,color='k')\n\n\nplt.show()\npltname = 'Anisotropy_S1'\nfig.savefig('../plots/'+pltname+'.pdf',bbox_inches='tight')\nfig.savefig('../plots/plots_png/'+pltname+'.png',bbox_inches='tight')", "/Users/ciaranohare/anaconda/lib/python3.7/site-packages/ipykernel_launcher.py:7: RuntimeWarning: divide by zero encountered in true_divide\n import sys\n\n/Users/ciaranohare/anaconda/lib/python3.7/site-packages/ipykernel_launcher.py:7: RuntimeWarning: invalid value encountered in true_divide\n import sys\n\n/Users/ciaranohare/anaconda/lib/python3.7/site-packages/ipykernel_launcher.py:8: RuntimeWarning: invalid value encountered in true_divide\n \n\n/Users/ciaranohare/anaconda/lib/python3.7/site-packages/ipykernel_launcher.py:11: RuntimeWarning: divide by zero encountered in true_divide\n # This is added back by InteractiveShellApp.init_path()\n\n/Users/ciaranohare/anaconda/lib/python3.7/site-packages/ipykernel_launcher.py:11: RuntimeWarning: invalid value encountered in true_divide\n # This is added back by InteractiveShellApp.init_path()\n\n/Users/ciaranohare/anaconda/lib/python3.7/site-packages/ipykernel_launcher.py:12: RuntimeWarning: invalid value encountered in true_divide\n if sys.path[0] == '':\n\n" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "lossess = [nn.L1Loss,nn.MSELoss,torch.nn.HingeEmbeddingLoss,torch.nn.MarginRankingLoss,torch.nn.TripletMarginLossnn.BCELoss]\nfor criterion in lossess:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh()).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.25)\n criterion = criterion()\n wandb.init(project=PROJECT_NAME,name=f'criterion-{criterion}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "import pandas as pd", "_____no_output_____" ], [ "data = pd.read_csv('./data.csv')", "_____no_output_____" ], [ "X,y = data.drop('target',axis=1),data['target']", "_____no_output_____" ], [ "from sklearn.model_selection import train_test_split", "_____no_output_____" ], [ "X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.25)", "_____no_output_____" ], [ "import torch\nimport torch.nn as nn", "_____no_output_____" ], [ "import numpy as np", "_____no_output_____" ], [ "X_train = torch.from_numpy(np.array(X_train).astype(np.float32))\ny_train = torch.from_numpy(np.array(y_train).astype(np.float32))\nX_test = torch.from_numpy(np.array(X_test).astype(np.float32))\ny_test = torch.from_numpy(np.array(y_test).astype(np.float32))", "_____no_output_____" ], [ "X_train.shape", "torch.Size([227, 13])" ], [ "X_test.shape", "torch.Size([76, 13])" ], [ "y_train.shape", "torch.Size([227])" ], [ "y_test.shape", "torch.Size([76])" ], [ "import torch.nn.functional as F", "_____no_output_____" ], [ "class Test_Model(nn.Module):\n def __init__(self):\n super().__init__()\n self.fc1 = nn.Linear(13,64)\n self.fc2 = nn.Linear(64,128)\n self.fc3 = nn.Linear(128,256)\n self.fc4 = nn.Linear(256,512)\n self.fc5 = nn.Linear(512,1024)\n self.fc6 = nn.Linear(1024,512)\n self.fc7 = nn.Linear(512,1)\n \n def forward(self,X):\n preds = self.fc1(X)\n preds = F.relu(preds)\n preds = self.fc2(preds)\n preds = F.relu(preds)\n preds = self.fc3(preds)\n preds = F.relu(preds)\n preds = self.fc4(preds)\n preds = F.relu(preds)\n preds = self.fc5(preds)\n preds = F.relu(preds)\n preds = self.fc6(preds)\n preds = F.relu(preds)\n preds = self.fc7(preds)\n return F.sigmoid(preds)", "_____no_output_____" ], [ "device = torch.device('cuda')", "_____no_output_____" ], [ "X_train = X_train.to(device)\ny_train = y_train.to(device)\nX_test = X_test.to(device)\ny_test = y_test.to(device)", "_____no_output_____" ], [ "PROJECT_NAME = 'Heart-Disease-UCI'", "_____no_output_____" ], [ "def get_loss(criterion,X,y,model):\n model.eval()\n with torch.no_grad():\n preds = model(X.float().to(device))\n preds = preds.view(len(preds),).to(device)\n y = y.view(len(y),).to(device)\n loss = criterion(preds,y)\n model.train()\n return loss.item()\ndef get_accuracy(preds,y):\n correct = 0\n total = 0\n for real,pred in zip(y_train,preds):\n if real == pred:\n correct += 1\n total += 1\n return round(correct/total,3)", "_____no_output_____" ], [ "import wandb", "_____no_output_____" ], [ "from tqdm import tqdm", "_____no_output_____" ], [ "EPOCHS = 212\n# EPOCHS = 100", "_____no_output_____" ], [ "# model = Test_Model().to(device)\n# optimizer = torch.optim.SGD(model.parameters(),lr=0.25)\n# criterion = nn.L1Loss()\n# wandb.init(project=PROJECT_NAME,name='baseline')\n# for _ in tqdm(range(EPOCHS)):\n# preds = model(X_train.float().to(device))\n# preds = preds.view(len(preds),)\n# preds.to(device)\n# loss = criterion(preds,y_train)\n# optimizer.zero_grad()\n# loss.backward()\n# optimizer.step()\n# wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(X_train,y_train,model),'val_accuracy':get_accuracy(X_test,y_test,model)})\n# wandb.finish()", "_____no_output_____" ], [ "# preds[:10]", "_____no_output_____" ], [ "# preds = torch.round(preds)", "_____no_output_____" ], [ "# correct = 0\n# total = 0\n# for real,pred in zip(y_train,preds):\n# if real == pred:\n# correct += 1\n# # total += 1", "_____no_output_____" ], [ "# round(correct/total,3)", "_____no_output_____" ], [ "## Testing Modelling", "_____no_output_____" ], [ "import torch\nimport torch.nn as nn", "_____no_output_____" ], [ "class Test_Model(nn.Module):\n def __init__(self,num_of_layers=1,activation=F.relu,input_shape=13,fc1_output=32,fc2_output=64,fc3_output=128,fc4_output=256,output_shape=1):\n super().__init__()\n self.num_of_layers = num_of_layers\n self.activation = activation\n self.fc1 = nn.Linear(input_shape,fc1_output)\n self.fc2 = nn.Linear(fc1_output,fc2_output)\n self.fc3 = nn.Linear(fc2_output,fc3_output)\n self.fc4 = nn.Linear(fc3_output,fc4_output)\n self.fc5 = nn.Linear(fc4_output,fc3_output)\n self.fc6 = nn.Linear(fc3_output,fc3_output)\n self.fc7 = nn.Linear(fc3_output,output_shape)\n \n def forward(self,X,activation=False):\n preds = self.fc1(X)\n if activation:\n preds = self.activation(preds)\n preds = self.fc2(preds)\n if activation:\n preds = self.activation(preds)\n preds = self.fc3(preds)\n if activation:\n preds = self.activation(preds)\n preds = self.fc4(preds)\n if activation:\n preds = self.activation(preds)\n preds = self.fc5(preds)\n if activation:\n preds = self.activation(preds)\n for _ in range(self.num_of_layers):\n preds = self.fc6(preds)\n if activation:\n preds = self.activation(preds)\n preds = self.fc7(preds)\n preds = F.sigmoid(preds)\n return preds", "_____no_output_____" ], [ "device = torch.device('cuda')", "_____no_output_____" ], [ "# preds = torch.round(preds)", "_____no_output_____" ], [ "# num_of_layers = 1\n# input_shape\n# fc1_output\n# fc2_output\n# fc3_output\n# fc4_output\n# output_shape\n# optimizer = torch.optim.SGD\n# criterion = \n# lr\n# activtion = nn.Tanh()", "_____no_output_____" ], [ "lossess = [nn.L1Loss,nn.MSELoss,torch.nn.HingeEmbeddingLoss,torch.nn.MarginRankingLoss,torch.nn.TripletMarginLossnn.BCELoss]\nfor criterion in lossess:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh()).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.25)\n criterion = criterion()\n wandb.init(project=PROJECT_NAME,name=f'criterion-{criterion}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "lossess = [nn.L1Loss,nn.MSELoss,torch.nn.HingeEmbeddingLoss,torch.nn.MarginRankingLoss,torch.nn.TripletMarginLossnn,torch.nn.BCELoss]\nfor criterion in lossess:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh()).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.25)\n criterion = criterion()\n wandb.init(project=PROJECT_NAME,name=f'criterion-{criterion}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "lossess = [nn.L1Loss,nn.MSELoss,torch.nn.HingeEmbeddingLoss,torch.nn.MarginRankingLoss,torch.nn.TripletMarginLoss,torch.nn.BCELoss]\nfor criterion in lossess:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh()).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.25)\n criterion = criterion()\n wandb.init(project=PROJECT_NAME,name=f'criterion-{criterion}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "# nn.L1Loss,nn.MSELoss,torch.nn.HingeEmbeddingLoss,\nlossess = [torch.nn.TripletMarginLoss,torch.nn.BCELoss]\nfor criterion in lossess:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh()).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.25)\n criterion = criterion()\n wandb.init(project=PROJECT_NAME,name=f'criterion-{criterion}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "# nn.L1Loss,nn.MSELoss,torch.nn.HingeEmbeddingLoss,\nlossess = [torch.nn.BCELoss]\nfor criterion in lossess:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh()).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.25)\n criterion = criterion()\n wandb.init(project=PROJECT_NAME,name=f'criterion-{criterion}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "lrs = [0.1,1.0,0.25,0.125,0.5,0.75,0.01,0.001,0.0001]\nfor lr in lrs:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh()).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=lr)\n criterion = nn.MSELoss()\n wandb.init(project=PROJECT_NAME,name=f'lr-{lr}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "fc1_outputs = [16,32,64,128,256]\nfor fc1_output in fc1_outputs:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh()fc1_outputs=fc1_output).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.125)\n criterion = nn.MSELoss()\n wandb.init(project=PROJECT_NAME,name=f'fc1_output-{fc1_output}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "fc1_outputs = [16,32,64,128,256]\nfor fc1_output in fc1_outputs:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh(),fc1_outputs=fc1_output).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.125)\n criterion = nn.MSELoss()\n wandb.init(project=PROJECT_NAME,name=f'fc1_output-{fc1_output}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "fc1_outputs = [16,32,64,128,256]\nfor fc1_output in fc1_outputs:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh(),fc1_output=fc1_output).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.125)\n criterion = nn.MSELoss()\n wandb.init(project=PROJECT_NAME,name=f'fc1_output-{fc1_output}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "fc2_outputs = [16,32,64,128,256,512]\nfor fc2_output in fc2_outputs:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh(),fc1_output=256,fc2_output=fc2_output).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.125)\n criterion = nn.MSELoss()\n wandb.init(project=PROJECT_NAME,name=f'fc2_output-{fc2_output}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "fc3_outputs = [16,32,64,128,256,512,1024]\nfor fc3_output in fc3_outputs:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh(),fc1_output=256,fc2_output=64,fc3_output=fc3_output).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.125)\n criterion = nn.MSELoss()\n wandb.init(project=PROJECT_NAME,name=f'fc3_output-{fc3_output}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ], [ "# num_of_layers = 1\n# fc1_output = 256\n# fc2_output = 64\n# fc3_output = 32\n# fc4_output = \n# optimizer = torch.optim.SGD\n# criterion = nn.MSELoss\n# lr = 0.125\n# activtion = nn.Tanh()", "_____no_output_____" ], [ "fc4_outputs = [16,32,64,128,256,512,1024,2048]\nfor fc4_output in fc4_outputs:\n model = Test_Model(num_of_layers=1,activation=nn.Tanh(),fc1_output=256,fc2_output=64,fc3_output=32,fc4_output=fc4_output).to(device)\n model.to(device)\n optimizer = torch.optim.SGD(model.parameters(),lr=0.125)\n criterion = nn.MSELoss()\n wandb.init(project=PROJECT_NAME,name=f'fc4_output-{fc4_output}')\n for _ in tqdm(range(212)):\n preds = model(X_train.float().to(device),True)\n preds = preds.view(len(preds),)\n preds.to(device)\n loss = criterion(preds,y_train)\n optimizer.zero_grad()\n loss.backward()\n optimizer.step()\n wandb.log({'loss':loss.item(),'val_loss':get_loss(criterion,X_test,y_test,model),'accuracy':get_accuracy(preds,y_train)})\n wandb.finish()", "_____no_output_____" ] ] ]

[ "code" ]

[ [ "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Example notebook for training a U-net deep learning network to predict tree cover", "_____no_output_____" ], [ "This notebook presents a toy example for training a deep learning architecture for semantic segmentation of satellite images using `eo-learn` and `keras`. The example showcases tree cover prediction over an area in Framce. The ground-truth data is retrieved from the [EU tree cover density (2015)](https://land.copernicus.eu/pan-european/high-resolution-layers/forests/view) through [Geopedia](http://www.geopedia.world/#T235_L2081_x449046.043261205_y6052157.300792162_s15_b17).\n\nThe workflow is as foolows:\n * input the area-of-interest (AOI)\n * split the AOI into small manageable eopatches\n * for each eopatch:\n * download RGB bands form Sentinel-2 L2A products using Sentinel-Hub for the 2017 year \n * retrieve corresponding ground-truth from Geopedia using a WMS request\n * compute the median values for the RGB bands over the time-interval\n * save to disk\n * select a 256x256 patch with corresponding ground-truth to be used for training/validating the model\n * train and validate a U-net\n \nThis example can easily be expanded to:\n * larger AOIs;\n * include more/different bands/indices, such as NDVI\n * include Sentinel-1 images (after harmonisation with Sentinel-2)\n \nThe notebook requires `Keras` with `tensorflow` back-end.", "_____no_output_____" ] ], [ [ "import os\nimport datetime \nfrom os import path as op\nimport itertools\n\nfrom eolearn.io import *\nfrom eolearn.core import EOTask, EOPatch, LinearWorkflow, FeatureType, SaveToDisk, OverwritePermission\nfrom sentinelhub import BBox, CRS, BBoxSplitter, MimeType, ServiceType\n\nfrom tqdm import tqdm_notebook as tqdm\nimport matplotlib.pyplot as plt\nimport numpy as np\nimport geopandas\n\nfrom sklearn.metrics import confusion_matrix\nfrom keras.preprocessing.image import ImageDataGenerator\nfrom keras import backend as K\nfrom keras.models import *\nfrom keras.layers import *\nfrom keras.optimizers import *\nfrom keras.utils.np_utils import to_categorical\nK.clear_session()", "Using TensorFlow backend.\n" ] ], [ [ "## 1. Set up workflow", "_____no_output_____" ] ], [ [ "# global image request parameters\ntime_interval = ('2017-01-01', '2017-12-31')\nimg_width = 256\nimg_height = 256\nmaxcc = 0.2", "_____no_output_____" ], [ "# get the AOI and split into bboxes\ncrs = CRS.UTM_31N\naoi = geopandas.read_file('../../example_data/eastern_france.geojson')\naoi = aoi.to_crs(crs=crs.pyproj_crs())\naoi_shape = aoi.geometry.values[-1]\n\nbbox_splitter = BBoxSplitter([aoi_shape], crs, (19, 10))", "_____no_output_____" ], [ "# set raster_value conversions for our Geopedia task\n# see more about how to do this here:\n\nraster_value = {\n '0%': (0, [0, 0, 0, 0]),\n '10%': (1, [163, 235, 153, 255]),\n '30%': (2, [119, 195, 118, 255]),\n '50%': (3, [85, 160, 89, 255]),\n '70%': (4, [58, 130, 64, 255]),\n '90%': (5, [36, 103, 44, 255])\n}", "_____no_output_____" ], [ "import matplotlib as mpl\n\ntree_cmap = mpl.colors.ListedColormap(['#F0F0F0', \n '#A2EB9B', \n '#77C277', \n '#539F5B', \n '#388141', \n '#226528'])\ntree_cmap.set_over('white')\ntree_cmap.set_under('white')\n\nbounds = np.arange(-0.5, 6, 1).tolist()\ntree_norm = mpl.colors.BoundaryNorm(bounds, tree_cmap.N)", "_____no_output_____" ], [ "# create a task for calculating a median pixel value\nclass MedianPixel(EOTask):\n \"\"\"\n The task returns a pixelwise median value from a time-series and stores the results in a \n timeless data array.\n \"\"\"\n def __init__(self, feature, feature_out):\n self.feature_type, self.feature_name = next(self._parse_features(feature)())\n self.feature_type_out, self.feature_name_out = next(self._parse_features(feature_out)())\n\n def execute(self, eopatch):\n eopatch.add_feature(self.feature_type_out, self.feature_name_out, \n np.median(eopatch[self.feature_type][self.feature_name], axis=0))\n return eopatch", "_____no_output_____" ], [ "# initialize tasks\n# task to get S2 L2A images\n\ninput_task = SentinelHubInputTask(data_source=DataSource.SENTINEL2_L2A, \n bands_feature=(FeatureType.DATA, 'BANDS'),\n resolution=10, \n maxcc=0.2, \n bands=['B04', 'B03', 'B02'], \n time_difference=datetime.timedelta(hours=2),\n additional_data=[(FeatureType.MASK, 'dataMask', 'IS_DATA')]\n )\ngeopedia_data = AddGeopediaFeature((FeatureType.MASK_TIMELESS, 'TREE_COVER'), \n layer='ttl2275', theme='QP', raster_value=raster_value)\n# task to compute median values\nget_median_pixel = MedianPixel((FeatureType.DATA, 'BANDS'), \n feature_out=(FeatureType.DATA_TIMELESS, 'MEDIAN_PIXEL'))\n# task to save to disk\nsave = SaveTask(op.join('data', 'eopatch'), \n overwrite_permission=OverwritePermission.OVERWRITE_PATCH, \n compress_level=2)", "_____no_output_____" ], [ "# initialize workflow\nworkflow = LinearWorkflow(input_task, geopedia_data, get_median_pixel, save)", "_____no_output_____" ], [ "# use a function to run this workflow on a single bbox\ndef execute_workflow(index):\n bbox = bbox_splitter.bbox_list[index]\n info = bbox_splitter.info_list[index]\n \n patch_name = 'eopatch_{0}_row-{1}_col-{2}'.format(index, \n info['index_x'], \n info['index_y'])\n \n results = workflow.execute({input_task:{'bbox':bbox, 'time_interval':time_interval},\n save:{'eopatch_folder':patch_name}\n })\n return list(results.values())[-1]\n del results ", "_____no_output_____" ] ], [ [ "Test workflow on an example patch and display", "_____no_output_____" ] ], [ [ "idx = 168\nexample_patch = execute_workflow(idx) ", "_____no_output_____" ], [ "example_patch", "_____no_output_____" ], [ "mp = example_patch.data_timeless['MEDIAN_PIXEL']\nplt.figure(figsize=(15,15))\nplt.imshow(2.5*mp)\ntc = example_patch.mask_timeless['TREE_COVER']\nplt.imshow(tc[...,0], vmin=0, vmax=5, alpha=.5, cmap=tree_cmap)\nplt.colorbar()", "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n" ] ], [ [ "## 2. Run workflow on all patches", "_____no_output_____" ] ], [ [ "# run over multiple bboxes\nsubset_idx = len(bbox_splitter.bbox_list)\nx_train_raw = np.empty((subset_idx, img_height, img_width, 3))\ny_train_raw = np.empty((subset_idx, img_height, img_width, 1))\npbar = tqdm(total=subset_idx)\nfor idx in range(0, subset_idx):\n patch = execute_workflow(idx)\n x_train_raw[idx] = patch.data_timeless['MEDIAN_PIXEL'][20:276,0:256,:]\n y_train_raw[idx] = patch.mask_timeless['TREE_COVER'][20:276,0:256,:]\n pbar.update(1)", "_____no_output_____" ] ], [ [ "## 3. Create training and validation data arrays", "_____no_output_____" ] ], [ [ "# data normalization and augmentation\nimg_mean = np.mean(x_train_raw, axis=(0, 1, 2))\nimg_std = np.std(x_train_raw, axis=(0, 1, 2))\nx_train_mean = x_train_raw - img_mean\nx_train = x_train_mean - img_std\n\ntrain_gen = ImageDataGenerator(\n horizontal_flip=True,\n vertical_flip=True,\n rotation_range=180)\n\ny_train = to_categorical(y_train_raw, len(raster_value))", "_____no_output_____" ] ], [ [ "## 4. Set up U-net model using Keras (tensorflow back-end)", "_____no_output_____" ] ], [ [ "# Model setup\n#from https://www.kaggle.com/lyakaap/weighing-boundary-pixels-loss-script-by-keras2\n# weight: weighted tensor(same shape with mask image)\ndef weighted_bce_loss(y_true, y_pred, weight):\n # avoiding overflow\n epsilon = 1e-7\n y_pred = K.clip(y_pred, epsilon, 1. - epsilon)\n logit_y_pred = K.log(y_pred / (1. - y_pred))\n\n # https://www.tensorflow.org/api_docs/python/tf/nn/weighted_cross_entropy_with_logits\n loss = (1. - y_true) * logit_y_pred + (1. + (weight - 1.) * y_true) * \\\n (K.log(1. + K.exp(-K.abs(logit_y_pred))) + K.maximum(-logit_y_pred, 0.))\n return K.sum(loss) / K.sum(weight)\n\ndef weighted_dice_loss(y_true, y_pred, weight):\n smooth = 1.\n w, m1, m2 = weight * weight, y_true, y_pred\n intersection = (m1 * m2)\n score = (2. * K.sum(w * intersection) + smooth) / (K.sum(w * m1) + K.sum(w * m2) + smooth)\n loss = 1. - K.sum(score)\n return loss\n\ndef weighted_bce_dice_loss(y_true, y_pred):\n y_true = K.cast(y_true, 'float32')\n y_pred = K.cast(y_pred, 'float32')\n # if we want to get same size of output, kernel size must be odd number\n averaged_mask = K.pool2d(\n y_true, pool_size=(11, 11), strides=(1, 1), padding='same', pool_mode='avg')\n border = K.cast(K.greater(averaged_mask, 0.005), 'float32') * K.cast(K.less(averaged_mask, 0.995), 'float32')\n weight = K.ones_like(averaged_mask)\n w0 = K.sum(weight)\n weight += border * 2\n w1 = K.sum(weight)\n weight *= (w0 / w1)\n loss = weighted_bce_loss(y_true, y_pred, weight) + \\\n weighted_dice_loss(y_true, y_pred, weight)\n return loss\n\ndef unet(input_size):\n inputs = Input(input_size)\n conv1 = Conv2D(64, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(inputs)\n conv1 = Conv2D(64, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(conv1)\n pool1 = MaxPooling2D(pool_size=(2, 2))(conv1)\n conv2 = Conv2D(128, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(pool1)\n conv2 = Conv2D(128, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(conv2)\n pool2 = MaxPooling2D(pool_size=(2, 2))(conv2)\n conv3 = Conv2D(256, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(pool2)\n conv3 = Conv2D(256, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(conv3)\n pool3 = MaxPooling2D(pool_size=(2, 2))(conv3)\n conv4 = Conv2D(512, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(pool3)\n conv4 = Conv2D(512, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(conv4)\n drop4 = Dropout(0.5)(conv4)\n pool4 = MaxPooling2D(pool_size=(2, 2))(drop4)\n\n conv5 = Conv2D(1024, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(pool4)\n conv5 = Conv2D(1024, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(conv5)\n drop5 = Dropout(0.5)(conv5)\n\n up6 = Conv2D(512, 2, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(UpSampling2D(size = (2,2))(drop5))\n merge6 = concatenate([drop4,up6])\n conv6 = Conv2D(512, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(merge6)\n conv6 = Conv2D(512, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(conv6)\n\n up7 = Conv2D(256, 2, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(UpSampling2D(size = (2,2))(conv6))\n merge7 = concatenate([conv3,up7])\n conv7 = Conv2D(256, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(merge7)\n conv7 = Conv2D(256, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(conv7)\n\n up8 = Conv2D(128, 2, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(UpSampling2D(size = (2,2))(conv7))\n merge8 = concatenate([conv2,up8])\n conv8 = Conv2D(128, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(merge8)\n conv8 = Conv2D(128, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(conv8)\n\n up9 = Conv2D(64, 2, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(UpSampling2D(size = (2,2))(conv8))\n merge9 = concatenate([conv1,up9])\n conv9 = Conv2D(64, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(merge9)\n conv9 = Conv2D(64, 3, activation = 'relu', padding = 'same', \n kernel_initializer = 'he_normal')(conv9)\n conv10 = Conv2D(len(raster_value), 1, activation = 'softmax')(conv9)\n\n model = Model(inputs = inputs, outputs = conv10)\n\n model.compile(optimizer = Adam(lr = 1e-4), \n loss = weighted_bce_dice_loss, \n metrics = ['accuracy'])\n\n return model\n\nmodel = unet(input_size=(256, 256, 3))", "_____no_output_____" ] ], [ [ "## 5. Train the model", "_____no_output_____" ] ], [ [ "# Fit the model\nbatch_size = 16\nmodel.fit_generator(\n train_gen.flow(x_train, y_train, batch_size=batch_size),\n steps_per_epoch=len(x_train),\n epochs=20,\n verbose=1)\nmodel.save(op.join('model.h5'))", "Epoch 1/20\n190/190 [==============================] - 419s 2s/step - loss: 0.9654 - acc: 0.6208\nEpoch 2/20\n190/190 [==============================] - 394s 2s/step - loss: 0.9242 - acc: 0.6460\nEpoch 3/20\n190/190 [==============================] - 394s 2s/step - loss: 0.9126 - acc: 0.6502\nEpoch 4/20\n190/190 [==============================] - 393s 2s/step - loss: 0.9059 - acc: 0.6540\nEpoch 5/20\n190/190 [==============================] - 393s 2s/step - loss: 0.9051 - acc: 0.6575\nEpoch 6/20\n190/190 [==============================] - 393s 2s/step - loss: 0.8940 - acc: 0.6620\nEpoch 7/20\n190/190 [==============================] - 394s 2s/step - loss: 0.8896 - acc: 0.6654\nEpoch 8/20\n190/190 [==============================] - 393s 2s/step - loss: 0.8870 - acc: 0.6668\nEpoch 9/20\n190/190 [==============================] - 393s 2s/step - loss: 0.8844 - acc: 0.6671\nEpoch 10/20\n190/190 [==============================] - 393s 2s/step - loss: 0.8781 - acc: 0.6708\nEpoch 11/20\n190/190 [==============================] - 393s 2s/step - loss: 0.8712 - acc: 0.6766\nEpoch 12/20\n190/190 [==============================] - 393s 2s/step - loss: 0.8671 - acc: 0.6801\nEpoch 13/20\n190/190 [==============================] - 394s 2s/step - loss: 0.8606 - acc: 0.6827\nEpoch 14/20\n190/190 [==============================] - 393s 2s/step - loss: 0.8489 - acc: 0.6919\nEpoch 15/20\n190/190 [==============================] - 394s 2s/step - loss: 0.8393 - acc: 0.6982\nEpoch 16/20\n190/190 [==============================] - 394s 2s/step - loss: 0.8279 - acc: 0.7063\nEpoch 17/20\n190/190 [==============================] - 394s 2s/step - loss: 0.8160 - acc: 0.7129\nEpoch 18/20\n190/190 [==============================] - 394s 2s/step - loss: 0.8020 - acc: 0.7233\nEpoch 19/20\n190/190 [==============================] - 394s 2s/step - loss: 0.7849 - acc: 0.7342\nEpoch 20/20\n190/190 [==============================] - 394s 2s/step - loss: 0.7781 - acc: 0.7397\n" ] ], [ [ "## 6. Validate model and show some results", "_____no_output_____" ] ], [ [ "# plot one example (image, label, prediction)\nidx = 4\np = np.argmax(model.predict(np.array([x_train[idx]])), axis=3)\nfig = plt.figure(figsize=(12,4))\nax1 = fig.add_subplot(1,3,1)\nax1.imshow(x_train_raw[idx])\nax2 = fig.add_subplot(1,3,2)\nax2.imshow(y_train_raw[idx][:,:,0])\nax3 = fig.add_subplot(1,3,3)\nax3.imshow(p[0])", "_____no_output_____" ], [ "def plot_confusion_matrix(cm, classes,\n normalize=False,\n title='Confusion matrix',\n cmap=plt.cm.Blues):\n \"\"\"\n This function prints and plots the confusion matrix.\n Normalization can be applied by setting `normalize=True`.\n \"\"\"\n if normalize:\n cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n print(\"Normalized confusion matrix\")\n else:\n print('Confusion matrix, without normalization')\n\n print(cm)\n\n plt.imshow(cm, interpolation='nearest', cmap=cmap)\n plt.title(title)\n plt.colorbar()\n tick_marks = np.arange(len(classes))\n plt.xticks(tick_marks, classes, rotation=45)\n plt.yticks(tick_marks, classes)\n\n fmt = '.2f' if normalize else 'd'\n thresh = cm.max() / 2.\n for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n plt.text(j, i, format(cm[i, j], fmt),\n horizontalalignment=\"center\",\n color=\"white\" if cm[i, j] > thresh else \"black\")\n\n plt.ylabel('True label')\n plt.xlabel('Predicted label')\n plt.tight_layout()", "_____no_output_____" ], [ "# show image confusion matrix\npredictions = np.argmax(model.predict(x_train), axis=3)\ncnf_matrix = confusion_matrix(y_train_raw.reshape(len(y_train_raw) * 256 * 256, 1), \n predictions.reshape(len(predictions) * 256 * 256, 1))\nplot_confusion_matrix(cnf_matrix, raster_value.keys(), normalize=True)", "Normalized confusion matrix\n[[0.93412552 0. 0. 0. 0.01624412 0.04963036]\n [0.75458006 0. 0. 0. 0.08682321 0.15859672]\n [0.73890185 0. 0. 0. 0.1051384 0.15595975]\n [0.6504189 0. 0. 0. 0.1332155 0.2163656 ]\n [0.36706531 0. 0. 0. 0.18843914 0.44449555]\n [0.1816605 0. 0. 0. 0.05291638 0.76542312]]\n" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code" ], [ "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "import scipy\nfrom scipy import ndimage\nimport scipy.misc\nimport matplotlib.pyplot as plt\nimport numpy as np\nfrom PIL import Image\nfrom scipy.signal import find_peaks\nfrom skimage.morphology import skeletonize\nfrom skimage import util \nfrom skimage import measure\nfrom skimage import filters\nfrom skimage import data\nfrom skimage import exposure \nfrom skimage import io\nimport pandas as pd\nimport os\nimport os.path\nfrom datetime import datetime\nimport seaborn as sns", "/opt/anaconda3/lib/python3.7/site-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n import pandas.util.testing as tm\n" ], [ "def autophagosome_size(img):\n im= np.array(img)!=2\n mask, number_of_objects = ndimage.label(im)\n all_labels = measure.label(mask)\n props = measure.regionprops(all_labels) \n area=[] \n for i in range (0,number_of_objects): \n if props[i].area > 2 and props[i].area < 1500:\n area=np.append(area,props[i].area)\n area=np.reshape(area, (-1, 1))\n ### size thresholding can be inroduced here\n if number_of_objects is not 0: return number_of_objects,area.mean(),area\n if number_of_objects is 0: return number_of_objects,number_of_objects,area", "_____no_output_____" ] ], [ [ "## Measure autophagosome properties", "_____no_output_____" ], [ "area=[]\nfor i in range (0,number_of_objects):\n area=np.append(area,props[i].area)", "_____no_output_____" ] ], [ [ "experiments = os.listdir(os. getcwd())\nfor item in experiments:\n if 'raph' not in item : experiments.remove(item)\nexperiments.remove('.ipynb_checkpoints')\n#experiments.remove('Statistical_Analysis.ipynb')\nexperiments.remove('Measure_and _Plot.ipynb')\nexperiments.remove('train')\nexperiments.remove('Results')\nexperiments", "_____no_output_____" ], [ "images_dir = 'train/label/'\noutput_csv_dir = 'Results/'\nos.makedirs(output_csv_dir, exist_ok=True)\nresults_all = pd.DataFrame(columns=['experiment','condition','image_name','number_of_objects','mean_area'])\n\nfor experiment in experiments:\n conditions = os.listdir(experiment)\n if '.DS_Store' in conditions: conditions.remove('.DS_Store')\n if 'Screenshot_1.png' in conditions: conditions.remove('Screenshot_1.png')\n pooled_cell_sizes_expt=pd.DataFrame()\n mean_cell_sizes_expt=pd.DataFrame()###\n mean_cell_numbers_expt=pd.DataFrame()\n for condition in conditions:\n data_path = str(experiment+'/'+condition)\n images = os.listdir(data_path)\n if '.DS_Store' in images: images.remove('.DS_Store')\n if '0_tif_RGB' in images: images.remove('0_tif_RGB')\n if '200331_Figure_Atg8a_Chloroquine.jpg' in images: images.remove('200331_Figure_Atg8a_Chloroquine.jpg')\n pooled_cell_sizes =[]\n mean_cell_sizes=[]###\n mean_cell_numbers=[]\n for image_name in images:\n file_name = str(experiment)+'_'+str(condition)+'_'+str(image_name)#+'_Simple_Segmentation'\n if file_name.endswith('.tif'):file_name = file_name[:-4].__add__('_Simple Segmentation.tif')\n elif file_name.endswith('.tiff'):file_name = file_name[:-5].__add__('_Simple Segmentation.tif')\n image = io.imread(os.path.join(images_dir, file_name), plugin='pil')\n \n number_of_objects,mean_area,area_all_objects = autophagosome_size(image)\n \n mean_cell_sizes=np.append(mean_cell_sizes,mean_area)###\n mean_cell_numbers=np.append(mean_cell_numbers,number_of_objects)\n pooled_cell_sizes=np.append(pooled_cell_sizes,area_all_objects)\n results_all = results_all.append({'experiment': str(experiment), 'condition':str(condition), 'image_name': str(image_name), 'number_of_objects':number_of_objects,'mean_area': mean_area},ignore_index=True)\n\n pooled_cell_sizes_data = np.reshape(pooled_cell_sizes, (-1, 1))\n pooled_cell_sizes_df = pd.DataFrame(data=pooled_cell_sizes_data, index=None, columns=[str(condition)])\n pooled_cell_sizes_expt = pd.concat([pooled_cell_sizes_df,pooled_cell_sizes_expt], axis=1,join='outer')\n \n mean_cell_numbers_data = np.reshape(mean_cell_numbers, (-1, 1))\n mean_cell_numbers_df = pd.DataFrame(data=mean_cell_numbers_data, index=None, columns=[str(condition)])\n mean_cell_numbers_expt = pd.concat([mean_cell_numbers_df,mean_cell_numbers_expt], axis=1,join='outer')#### \n \n mean_cell_sizes_data = np.reshape(mean_cell_sizes, (-1, 1))\n mean_cell_sizes_df = pd.DataFrame(data=mean_cell_sizes_data, index=None, columns=[str(condition)])\n mean_cell_sizes_expt = pd.concat([mean_cell_sizes_df,mean_cell_sizes_expt], axis=1,join='outer')#### \n \n pooled_cell_sizes_expt.to_csv(output_csv_dir+experiment+'_pooled_cell_sizes.csv', sep=';', decimal=',')\n mean_cell_sizes_expt.to_csv(output_csv_dir+experiment+'_mean_cell_sizes.csv', sep=';', decimal=',')####\n mean_cell_numbers_expt.to_csv(output_csv_dir+experiment+'_mean_cell_numbers.csv', sep=';', decimal=',')\n print(experiment+' DONE')\nresults_all.to_csv(output_csv_dir+'results_all_'+str(datetime.now())+'.csv', sep=';', decimal=',')\nresults_all.to_csv(output_csv_dir+'results_all.csv', sep=';', decimal=',')\nprint('ALL DONE')", "Graph10_BoiPy__60xWater DONE\nGraph11_ER__60xWater DONE\nGraph12_Golgi__60xWater DONE\nGraph14_Atg8a_Epistase_time_Of_Woud_Healing__40x DONE\nGraph15_Atg8a_Insulin_Foxo_time_Of_Woud_Healing__40x_and_60x DONE\nGraph16_Atg8a_Foxo_TM_time_Of_Woud_Healing__40xOil DONE\nGraph17_Atg8a_time_Of_Woud_Healing__40xOil DONE\nGraph1_Geraf2__Atg8a__40xOil_rest_of_data DONE\n" ], [ "results_all=pd.read_csv(output_csv_dir+'results_all_2020-08-30-2.csv', sep=';', decimal=',')\nresults_all.head()", "_____no_output_____" ] ], [ [ "## PLOT NUMBER OF OBJECTS", "_____no_output_____" ] ], [ [ "\nimport seaborn as sns\nfor experiment in experiments:\n experiment_data = results_all[results_all['experiment']==experiment] \n fig,ax = plt.subplots()\n ax.spines['right'].set_visible(False)\n ax.spines['top'].set_visible(False)\n ax = sns.boxplot(x=\"condition\", y=\"number_of_objects\", data=experiment_data)\n ax = sns.swarmplot(x=\"condition\", y=\"number_of_objects\", data=experiment_data, color=\".25\")\n plt.xticks(rotation=90)\n plt.title(experiment)\n plt.savefig(output_csv_dir+experiment+'_number_of_objects.png',bbox_inches='tight')\n plt.show()", "_____no_output_____" ] ], [ [ "## PLOT SIZE OF OBJECTS/ MEAN_AREA", "_____no_output_____" ] ], [ [ "\n\nfor experiment in experiments:\n experiment_data = results_all[results_all['experiment']==experiment] \n fig,ax = plt.subplots()\n ax.spines['right'].set_visible(False)\n ax.spines['top'].set_visible(False)\n ax = sns.boxplot(x=\"condition\", y=\"mean_area\", data=experiment_data)\n ax = sns.swarmplot(x=\"condition\", y=\"mean_area\", data=experiment_data, color=\".25\")\n plt.xticks(rotation=90)\n plt.ylabel('mean_area ($\\mu$m$^{2}$)')\n plt.title(experiment)\n plt.savefig(output_csv_dir+experiment+'_mean_area.png',bbox_inches='tight')\n plt.show()", "_____no_output_____" ] ], [ [ "## PLOT POOLED SIZE OF OBJECTS/ MEAN_AREA", "_____no_output_____" ] ], [ [ "output_csv_dir = 'Results/'\nfor experiment in experiments:\n df=pd.read_csv(output_csv_dir+experiment+'_pooled_cell_sizes.csv', sep=';', decimal=',')\n df = df.drop(columns=['Unnamed: 0'])\n df = df.sort_index(axis=1)\n \n \n if '60x' in str(experiment):df=df*0.1111\n if '40x' in str(experiment):df=df*0.1626\n \n fig,ax = plt.subplots()\n ax.spines['right'].set_visible(False)\n ax.spines['top'].set_visible(False)\n ax = sns.stripplot(data=df, jitter=0.3,size=2)\n plt.xticks(rotation=90)\n plt.ylabel('spot size ($\\mu$m$^{2}$)')\n plt.ylim(-10,250)\n plt.title(experiment)\n plt.axhline(y=15,color='k')\n plt.savefig(output_csv_dir+experiment+'_pooled_cell_sizes.png',bbox_inches='tight')\n plt.show()", "_____no_output_____" ], [ "df.head()", "_____no_output_____" ], [ "def get_concat_h_multi_resize(im_list, resample=Image.BICUBIC):\n min_height = min(im.height for im in im_list)\n im_list_resize = [im.resize((int(im.width * min_height / im.height), min_height),resample=resample)\n for im in im_list]\n total_width = sum(im.width for im in im_list_resize)\n dst = Image.new('RGB', (total_width, min_height))\n pos_x = 0\n for im in im_list_resize:\n dst.paste(im, (pos_x, 0))\n pos_x += im.width\n return dst", "_____no_output_____" ], [ "## CONCATANATE GRAPHS\n\nplot_dir = 'Results/'\nfor experiment in experiments:\n print(experiment)\n im1 = Image.open(os.path.join(plot_dir, experiment+'_number_of_objects.png'))\n im2 = Image.open(os.path.join(plot_dir, experiment+'_mean_area.png'))\n im3 = Image.open(os.path.join(plot_dir, experiment+'_pooled_cell_sizes.png'))\n get_concat_h_multi_resize([im1, im2, im3]).save('Results/'+experiment+'_concat.jpg')", "Graph10_BoiPy__60xWater\nGraph11_ER__60xWater\nGraph12_Golgi__60xWater\nGraph14_Atg8a_Epistase_time_Of_Woud_Healing__40x\nGraph15_Atg8a_Insulin_Foxo_time_Of_Woud_Healing__40x_and_60x\nGraph16_Atg8a_Foxo_TM_time_Of_Woud_Healing__40xOil\nGraph17_Atg8a_time_Of_Woud_Healing__40xOil\nGraph1_Geraf2__Atg8a__40xOil_rest_of_data\nGraph3_Atg8a_Epistase__40xOil\nGraph4_Graph5_Atg8a_Chloroquine_you.have.all.images\nGraph6_Graph7_LAMP1.GFP__60xWater\nGraph8_Graph9_GFP.LAMP1__60xWater\n" ] ] ]

[ "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "code", "code" ], [ "markdown", "markdown" ], [ "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Инициализация", "_____no_output_____" ] ], [ [ "#@markdown - **Монтирование GoogleDrive** \nfrom google.colab import drive\ndrive.mount('GoogleDrive')", "_____no_output_____" ], [ "# #@markdown - **Размонтирование**\n# !fusermount -u GoogleDrive", "_____no_output_____" ] ], [ [ "# Область кодов", "_____no_output_____" ] ], [ [ "#@title Сверточные нейронные сети { display-mode: \"both\" }\n# В программе используется API в TensorFlow для реализации двухслойных сверточных нейронных сетей\n# coding: utf-8\nimport tensorflow.examples.tutorials.mnist.input_data as input_data\nimport tensorflow as tf\nimport matplotlib.pyplot as plt\nimport os", "_____no_output_____" ], [ "#@markdown - **Определение функций инициализации**\ndef glorot_init(shape, name):\n initial = tf.truncated_normal(shape=shape, stddev=1. / tf.sqrt(shape[0] / 2.))\n return tf.Variable(initial, name=name)\n\ndef bias_init(shape, name):\n initial = tf.constant(0.1, shape=shape)\n return tf.Variable(initial, name=name)\n\ndef conv2d(x, W):\n return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')\n\ndef max_pool_2x2(x):\n return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],\n strides=[1, 2, 2, 1], padding='SAME')", "_____no_output_____" ], [ "#@markdown - **Настройка гиперпараметров**\nmnist = input_data.read_data_sets('sample_data/MNIST_data', one_hot=True)\nnum_epochs = 12000 #@param {type: \"integer\"}\nbatch_size = 196 #@param {type: \"integer\"}\nlearning_rate = 8e-4 #@param {type: \"number\"}\n\ndir_path = 'GoogleDrive/My Drive/Colab Notebooks'\nevent_path = os.path.join(dir_path, 'Tensorboard')\ncheckpoint_path = os.path.join(dir_path, 'Checkpoints')\n", "WARNING:tensorflow:From <ipython-input-5-1291ae95062c>:1: read_data_sets (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\nInstructions for updating:\nPlease use alternatives such as official/mnist/dataset.py from tensorflow/models.\nWARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/contrib/learn/python/learn/datasets/mnist.py:260: maybe_download (from tensorflow.contrib.learn.python.learn.datasets.base) is deprecated and will be removed in a future version.\nInstructions for updating:\nPlease write your own downloading logic.\nWARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/contrib/learn/python/learn/datasets/mnist.py:262: extract_images (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\nInstructions for updating:\nPlease use tf.data to implement this functionality.\nExtracting sample_data/MNIST_data/train-images-idx3-ubyte.gz\nWARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/contrib/learn/python/learn/datasets/mnist.py:267: extract_labels (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\nInstructions for updating:\nPlease use tf.data to implement this functionality.\nExtracting sample_data/MNIST_data/train-labels-idx1-ubyte.gz\nWARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/contrib/learn/python/learn/datasets/mnist.py:110: dense_to_one_hot (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\nInstructions for updating:\nPlease use tf.one_hot on tensors.\nExtracting sample_data/MNIST_data/t10k-images-idx3-ubyte.gz\nExtracting sample_data/MNIST_data/t10k-labels-idx1-ubyte.gz\nWARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/contrib/learn/python/learn/datasets/mnist.py:290: DataSet.__init__ (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\nInstructions for updating:\nPlease use alternatives such as official/mnist/dataset.py from tensorflow/models.\n" ], [ "#@markdown - **Создание graph**\ngraph = tf.Graph()\nwith graph.as_default():\n\n with tf.name_scope('Input'):\n x = tf.placeholder(tf.float32, shape=[None, 784], name='input_images')\n y_ = tf.placeholder(tf.float32, shape=[None, 10], name='labels')\n x_image = tf.reshape(x, [-1, 28, 28, 1])\n keep_prob = tf.placeholder(tf.float32)\n\n #@markdown - **Настройка сверточных слоев**\n # --------------conv1-----------------------------------\n with tf.name_scope('Conv1'):\n with tf.name_scope('weights_conv1'):\n W_conv1 = glorot_init([3, 3, 1, 64], 'w_conv1')\n with tf.name_scope('bias_covn1'):\n b_conv1 = bias_init([64], 'b_conv1') \n h_conv1 = tf.nn.relu(conv2d(x_image, W_conv1) + b_conv1)\n with tf.name_scope('features_conv1'):\n h_pool1 = max_pool_2x2(h_conv1)\n\n # --------------conv2-----------------------------------\n with tf.name_scope('Conv2'):\n W_conv2 = glorot_init([3, 3, 64, 128], 'w_conv2')\n b_conv2 = bias_init([128], 'b_conv2')\n h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)\n h_pool2 = max_pool_2x2(h_conv2)\n \n #@markdown - **Настройка полносвязных слоев**\n # --------------fc--------------------------------------\n h_pool2_flat = tf.layers.flatten(h_pool2)\n num_f = h_pool2_flat.get_shape().as_list()[-1]\n with tf.name_scope('FC1'):\n W_fc1 = glorot_init([num_f, 128], 'w_fc1')\n b_fc1 = bias_init([128], 'b_fc1')\n h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)\n\n with tf.name_scope('Dropout'): \n h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)\n \n with tf.name_scope('FC2'):\n W_fc2 = glorot_init([128, 10], 'w_fc2')\n b_fc2 = bias_init([10], 'b_fc2')\n y_fc2 = tf.matmul(h_fc1_drop, W_fc2) + b_fc2\n\n with tf.name_scope('Loss'):\n y_out = tf.nn.softmax(y_fc2)\n# cross_entropy = -tf.reduce_mean(y_*tf.log(y_out + 1e-10))\n # # or like\n cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits_v2(labels=y_,\n logits=y_fc2))\n with tf.name_scope('Train'):\n train_step = tf.train.AdamOptimizer(learning_rate=learning_rate).minimize(cross_entropy)\n # # or like\n # optimizer = tf.train.AdamOptimizer(learning_rate=FLAGS.learning_rate)\n # grad_list = optimizer.compute_gradients(cross_entropy)\n # train_step = optimizer.apply_gradients(grad_list)\n\n with tf.name_scope('Accuracy'):\n correct_prediction = tf.equal(tf.argmax(y_out, 1), tf.argmax(y_, 1))\n accuracy = tf.reduce_mean(tf.cast(correct_prediction, \"float\"))", "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\nInstructions for updating:\nColocations handled automatically by placer.\nWARNING:tensorflow:From <ipython-input-6-95a34e4babde>:30: flatten (from tensorflow.python.layers.core) is deprecated and will be removed in a future version.\nInstructions for updating:\nUse keras.layers.flatten instead.\nWARNING:tensorflow:From <ipython-input-6-95a34e4babde>:38: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\nInstructions for updating:\nPlease use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n" ], [ "#@markdown - **Обучение сетей и сохранение моделей**\nwith tf.Session(graph=graph) as sess:\n\n sess.run(tf.global_variables_initializer())\n saver = tf.train.Saver(max_to_keep=3) # сохранить 3 модели\n max_acc = 101. # модели с более высокой точностью будут сохранены\n\n for epoch in range(num_epochs):\n batch = mnist.train.next_batch(batch_size)\n _, acc, loss = sess.run([train_step, accuracy, cross_entropy], feed_dict={x: batch[0], \n y_: batch[1], \n keep_prob: 0.5})\n step = epoch + 1\n if step % 1000 == 0:\n acc *= 100\n print_list = [step, loss, acc]\n print(\"Epoch: {0[0]}, cross_entropy: {0[1]:.4f}, accuracy on training data: {0[2]:.2f}%,\".format(print_list))\n test_acc, test_loss = sess.run([accuracy, cross_entropy], feed_dict={x: mnist.test.images, \n y_: mnist.test.labels, \n keep_prob: 1.0})\n test_acc *= 100\n print_list = [test_loss, test_acc]\n print(' '*12, 'cross_entropy: {0[0]:.4f}, accuracy on testing data: {0[1]:.2f}%.'.format(print_list))\n print('\\n')\n \n if (acc > max_acc) & (step > 3999):\n max_acc = acc\n saver.save(sess, os.path.join(checkpoint_path, 'f_map.ckpt'), global_step=step)\n\n test_image, test_label = mnist.test.images[100, :].reshape((1, -1)), mnist.test.labels[100, :].reshape((1, -1))\n features1, features2 = sess.run([h_pool1, h_pool2], feed_dict={x: test_image, y_: test_label, keep_prob: 1.0})", "Epoch: 1000, cross_entropy: 0.6997, accuracy on training data: 72.45%,\n cross_entropy: 0.1718, accuracy on testing data: 96.03%.\n\n\nEpoch: 2000, cross_entropy: 0.4519, accuracy on training data: 82.14%,\n cross_entropy: 0.0626, accuracy on testing data: 98.28%.\n\n\nEpoch: 3000, cross_entropy: 0.2301, accuracy on training data: 92.35%,\n cross_entropy: 0.0548, accuracy on testing data: 98.38%.\n\n\nEpoch: 4000, cross_entropy: 0.1407, accuracy on training data: 95.41%,\n cross_entropy: 0.0475, accuracy on testing data: 98.66%.\n\n\nEpoch: 5000, cross_entropy: 0.1878, accuracy on training data: 92.86%,\n cross_entropy: 0.0448, accuracy on testing data: 98.80%.\n\n\nEpoch: 6000, cross_entropy: 0.1025, accuracy on training data: 95.92%,\n cross_entropy: 0.0386, accuracy on testing data: 98.97%.\n\n\nEpoch: 7000, cross_entropy: 0.0510, accuracy on training data: 98.47%,\n cross_entropy: 0.0383, accuracy on testing data: 98.98%.\n\n\nEpoch: 8000, cross_entropy: 0.0281, accuracy on training data: 98.98%,\n cross_entropy: 0.0417, accuracy on testing data: 99.00%.\n\n\nEpoch: 9000, cross_entropy: 0.0042, accuracy on training data: 100.00%,\n cross_entropy: 0.0426, accuracy on testing data: 99.03%.\n\n\nEpoch: 10000, cross_entropy: 0.0260, accuracy on training data: 98.47%,\n cross_entropy: 0.0407, accuracy on testing data: 98.97%.\n\n\nEpoch: 11000, cross_entropy: 0.0075, accuracy on training data: 100.00%,\n cross_entropy: 0.0419, accuracy on testing data: 99.08%.\n\n\nEpoch: 12000, cross_entropy: 0.0324, accuracy on training data: 98.98%,\n cross_entropy: 0.0394, accuracy on testing data: 99.14%.\n\n\n" ], [ "#@markdown - **Восстановление сохраненной модели**\n# with tf.Session() as sess:\n # model_path = 'GoogleDrive/My Drive/Colab Notebooks/Tensorboard/f_map.ckpt-241'\n # saver.restore(sess, model_path)\n # acc, loss = sess.run([accuracy, cross_entropy], feed_dict={x: batch[0], y_: batch[1], keep_prob: 1.0})\n # print('Accuracy is %.2f.' %(acc))\n# sess.close()", "_____no_output_____" ], [ "#@markdown - **Представление feature map первого сверточного слоя**\nfeatures_map = features1.reshape((14, 14, 64))\nnum_map = range(features_map.shape[-1])\nfig, AX = plt.subplots(nrows=4, ncols=8)\nfig.set_size_inches(w=14, h=7)\nfig.subplots_adjust(wspace=.2, hspace=.2)\ntry:\n for index, ax in enumerate(AX.flatten()):\n ax.imshow(features_map[:, :, index], 'gray')\n ax.set_xticks([]), ax.set_yticks([])\nexcept IndexError:\n pass\nplt.show()", "_____no_output_____" ], [ "#@markdown - **Представление feature map второго сверточного слоя**\nfeatures_map = features2.reshape((7, 7, 128))\nnum_map = range(features_map.shape[-1])\nfig, AX = plt.subplots(nrows=4, ncols=8)\nfig.set_size_inches(w=14, h=7)\nfig.subplots_adjust(wspace=.2, hspace=.2)\ntry:\n for index, ax in enumerate(AX.flatten()):\n ax.imshow(features_map[:, :, index], 'gray')\n ax.set_xticks([]), ax.set_yticks([])\nexcept IndexError:\n pass\nplt.show()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code", "code", "code", "code", "code", "code", "code", "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Predicting Flight Delays with sklearn\n\nIn this notebook, we will be using features we've prepared in PySpark to predict flight delays via regression and classification.", "_____no_output_____" ] ], [ [ "import sys, os, re\nsys.path.append(\"lib\")\nimport utils\n\nimport numpy as np\nimport sklearn\nimport iso8601\nimport datetime\nprint(\"Imports loaded...\")", "Imports loaded...\n" ] ], [ [ "## Load and Inspect our JSON Training Data", "_____no_output_____" ] ], [ [ "# Load and check the size of our training data. May take a minute.\nprint(\"Original JSON file size: {:,} Bytes\".format(os.path.getsize(\"../data/simple_flight_delay_features.jsonl\")))\n\ntraining_data = utils.read_json_lines_file('../data/simple_flight_delay_features.jsonl')\n\nprint(\"Training items: {:,}\".format(len(training_data))) # 5,714,008\nprint(\"Data loaded...\")", "Original JSON file size: 4,096 Bytes\n" ], [ "# Inspect a record before we alter them\nprint(\"Size of training data in RAM: {:,} Bytes\".format(sys.getsizeof(training_data))) # 50MB\nprint(training_data[0])", "Size of training data in RAM: 406,496 Bytes\n{'ArrDelay': -14.0, 'CRSArrTime': '2015-01-01T10:25:00.000Z', 'CRSDepTime': '2015-01-01T08:55:00.000Z', 'Carrier': 'AA', 'DayOfMonth': 1, 'DayOfWeek': 4, 'DayOfYear': 1, 'DepDelay': -4.0, 'Dest': 'DFW', 'Distance': 731.0, 'FlightDate': '2015-01-01T00:00:00.000Z', 'FlightNum': '1455', 'Origin': 'ATL'}\n" ] ], [ [ "## Sample our Data", "_____no_output_____" ] ], [ [ "# We need to sample our data to fit into RAM\ntraining_data = np.random.choice(training_data, 1000000) # 'Sample down to 1MM examples'\nprint(\"Sampled items: {:,} Bytes\".format(len(training_data)))\nprint(\"Data sampled...\")", "Sampled items: 1,000,000 Bytes\nData sampled...\n" ] ], [ [ "## Vectorize the Results (y)", "_____no_output_____" ] ], [ [ "# Separate our results from the rest of the data, vectorize and size up\nresults = [record['ArrDelay'] for record in training_data]\nresults_vector = np.array(results)\nprint(\"Results vectorized size: {:,}\".format(sys.getsizeof(results_vector))) # 45,712,160 bytes\nprint(\"Results vectorized...\")", "Results vectorized size: 8,000,096\nResults vectorized...\n" ] ], [ [ "## Prepare Training Data", "_____no_output_____" ] ], [ [ "# Remove the two delay fields and the flight date from our training data\nfor item in training_data:\n item.pop('ArrDelay', None)\n item.pop('FlightDate', None)\nprint(\"ArrDelay and FlightDate removed from training data...\")", "ArrDelay and FlightDate removed from training data...\n" ], [ "# Must convert datetime strings to unix times\nfor item in training_data:\n if isinstance(item['CRSArrTime'], str):\n dt = iso8601.parse_date(item['CRSArrTime'])\n unix_time = int(dt.timestamp())\n item['CRSArrTime'] = unix_time\n if isinstance(item['CRSDepTime'], str):\n dt = iso8601.parse_date(item['CRSDepTime'])\n unix_time = int(dt.timestamp())\n item['CRSDepTime'] = unix_time\nprint(\"CRSArr/DepTime converted to unix time...\")", "CRSArr/DepTime converted to unix time...\n" ] ], [ [ "## Vectorize Training Data with `DictVectorizer`", "_____no_output_____" ] ], [ [ "# Use DictVectorizer to convert feature dicts to vectors\nfrom sklearn.feature_extraction import DictVectorizer\n\nprint(\"Sampled dimensions: [{:,}]\".format(len(training_data)))\nvectorizer = DictVectorizer()\ntraining_vectors = vectorizer.fit_transform(training_data)\nprint(\"Size of DictVectorized vectors: {:,} Bytes\".format(training_vectors.data.nbytes))\nprint(\"Training data vectorized...\")", "_____no_output_____" ] ], [ [ "## Prepare Experiment by Splitting Data into Train/Test", "_____no_output_____" ] ], [ [ "from sklearn.model_selection import train_test_split\n\nX_train, X_test, y_train, y_test = train_test_split(\n training_vectors,\n results_vector,\n test_size=0.1,\n random_state=43\n)\nprint(X_train.shape, X_test.shape)\nprint(y_train.shape, y_test.shape)\nprint(\"Test train split performed...\")", "_____no_output_____" ] ], [ [ "## Train our Model(s) on our Training Data", "_____no_output_____" ] ], [ [ "# Train a regressor\nfrom sklearn.linear_model import LinearRegression\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import median_absolute_error, r2_score\nprint(\"Regressor library and metrics imported...\")", "_____no_output_____" ], [ "regressor = LinearRegression()\nprint(\"Regressor instantiated...\")", "_____no_output_____" ], [ "from sklearn.ensemble import GradientBoostingRegressor\n\nregressor = GradientBoostingRegressor\nprint(\"Swapped gradient boosting trees for linear regression!\")\n\n# Lets go back for now...\nregressor = LinearRegression()\nprint(\"Swapped back to linear regression!\")", "_____no_output_____" ], [ "regressor.fit(X_train, y_train)\nprint(\"Regressor fitted...\")", "_____no_output_____" ] ], [ [ "## Predict Using the Test Data", "_____no_output_____" ] ], [ [ "predicted = regressor.predict(X_test)\nprint(\"Predictions made for X_test...\")", "_____no_output_____" ] ], [ [ "## Evaluate and Visualize Model Accuracy", "_____no_output_____" ] ], [ [ "from sklearn.metrics import median_absolute_error, r2_score\n\n# Median absolute error is the median of all absolute differences between the target and the prediction.\n# Less is better, more indicates a high error between target and prediction.\nmedae = median_absolute_error(y_test, predicted)\nprint(\"Median absolute error: {:.3g}\".format(medae))\n\n# R2 score is the coefficient of determination. Ranges from 1-0, 1.0 is best, 0.0 is worst.\n# Measures how well future samples are likely to be predicted.\nr2 = r2_score(y_test, predicted)\nprint(\"r2 score: {:.3g}\".format(r2))", "_____no_output_____" ], [ "# Plot outputs\nimport matplotlib.pyplot as plt\n\n# Cleans up the appearance\nplt.rcdefaults()\n\nplt.scatter(\n y_test,\n predicted,\n color='blue',\n linewidth=1\n)\nplt.grid(True)\n\nplt.xticks()\nplt.yticks()\n\nplt.show()", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code", "code", "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code", "code" ] ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ "Apache-2.0" ]

[ [ [ "# Assignment-1", "_____no_output_____" ] ], [ [ "class BankAccount(object):\n def __init__(self, initial_balance=0):\n self.balance = initial_balance\n def deposit(self, amount):\n self.balance += amount\n def withdraw(self, amount):\n self.balance -= amount\n def overdrawn(self):\n return self.balance < 0\nmy_account = BankAccount(15)\nmy_account.withdraw(5)\nprint (my_account.balance)\n", "10\n" ] ], [ [ "# Assignment-2", "_____no_output_____" ] ], [ [ "import math \npi = math.pi \n \n\ndef volume(r, h): \n return (1 / 3) * pi * r * r * h \n \n\ndef surfacearea(r, s): \n return pi * r * s + pi * r * r \n \n\nradius = float(5) \nheight = float(12) \nslat_height = float(13) \nprint( \"Volume Of Cone : \", volume(radius, height) ) \nprint( \"Surface Area Of Cone : \", surfacearea(radius, slat_height) ) ", "Volume Of Cone : 314.15926535897927\nSurface Area Of Cone : 282.7433388230814\n" ] ] ]

[ "markdown", "code", "markdown", "code" ]

[ [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ] ]

[ "MIT" ]

[ "MIT" ]

[ "MIT" ]

[ [ [ "# Praktikum 12 | Pengolahan Citra", "_____no_output_____" ], [ "## Sharpness\nSharpness adalah proses untuk mendapatkan gambar yang lebih tajam. Proses sharpness ini memanfaatkan BSF (Band-Stop Filter) yang merupakan gabungan antara LPF (Low Pass Filter) dan HPF (High Pass Filter).", "_____no_output_____" ], [ "Fadhil Yori Hibatullah | 2103161037 | 2 D3 Teknik Informatika B\n\n---------------------", "_____no_output_____" ], [ "### Import Dependency", "_____no_output_____" ] ], [ [ "import imageio\nimport matplotlib.pyplot as plt\nimport numpy as np", "_____no_output_____" ] ], [ [ "### Load Image", "_____no_output_____" ] ], [ [ "imgNormal = imageio.imread(\"gambar4.jpg\")", "_____no_output_____" ] ], [ [ "### Show Image", "_____no_output_____" ] ], [ [ "plt.imshow(imgNormal)\nplt.title(\"Load Image\")\nplt.show()", "_____no_output_____" ] ], [ [ "---------------------", "_____no_output_____" ], [ "## To Grayscale", "_____no_output_____" ] ], [ [ "imgGrayscale = np.zeros((imgNormal.shape[0], imgNormal.shape[1], 3), dtype=np.uint8)\n\nfor y in range(0, imgNormal.shape[0]):\n for x in range(0, imgNormal.shape[1]):\n r = imgNormal[y][x][0]\n g = imgNormal[y][x][1]\n b = imgNormal[y][x][2]\n gr = ( int(r) + int(g) + int(b) ) / 3\n imgGrayscale[y][x] = (gr, gr, gr)\n \nplt.imshow(imgGrayscale)\nplt.title(\"Grayscale\")\nplt.show()", "_____no_output_____" ] ], [ [ "---------------------", "_____no_output_____" ], [ "## Sharpness Gray", "_____no_output_____" ] ], [ [ "imgSharpnessGray = np.zeros((imgNormal.shape[0], imgNormal.shape[1], 3), dtype=np.uint8)\n\nfor y in range(1, imgNormal.shape[0] - 1):\n for x in range(1, imgNormal.shape[1] - 1):\n x1 = int(imgGrayscale[y - 1][x - 1][0])\n x2 = int(imgGrayscale[y][x - 1][0])\n x3 = int(imgGrayscale[y + 1][x - 1][0])\n x4 = int(imgGrayscale[y - 1][x][0])\n x5 = int(imgGrayscale[y][x][0])\n x6 = int(imgGrayscale[y + 1][x][0])\n x7 = int(imgGrayscale[y - 1][x + 1][0])\n x8 = int(imgGrayscale[y][x + 1][0])\n x9 = int(imgGrayscale[y + 1][x + 1][0])\n xt1 = int((x1 + x2 + x3 + x4 + x5 + x6 + x7 + x8 + x9) / 9)\n xt2 = int(-x1 - (2 * x2) - x3 + x7 + (2 * x8) + x9)\n xt3 = int(-x1 - (2 * x4) - x7 + x3 + (2 * x6) + x9)\n xb = int(xt1 + xt2 + xt3)\n if xb < 0:\n xb = -xb\n if xb > 255:\n xb = 255\n imgSharpnessGray[y][x] = (xb, xb, xb)\n \nplt.imshow(imgSharpnessGray)\nplt.title(\"Sharpness Gray\")\nplt.show()", "_____no_output_____" ] ], [ [ "------------------", "_____no_output_____" ], [ "## Sharpness Gray (2:1 LPF:HPF)", "_____no_output_____" ] ], [ [ "imgSharpnessGrayL = np.zeros((imgNormal.shape[0], imgNormal.shape[1], 3), dtype=np.uint8)\n\nfor y in range(1, imgNormal.shape[0] - 1):\n for x in range(1, imgNormal.shape[1] - 1):\n x1 = int(imgGrayscale[y - 1][x - 1][0])\n x2 = int(imgGrayscale[y][x - 1][0])\n x3 = int(imgGrayscale[y + 1][x - 1][0])\n x4 = int(imgGrayscale[y - 1][x][0])\n x5 = int(imgGrayscale[y][x][0])\n x6 = int(imgGrayscale[y + 1][x][0])\n x7 = int(imgGrayscale[y - 1][x + 1][0])\n x8 = int(imgGrayscale[y][x + 1][0])\n x9 = int(imgGrayscale[y + 1][x + 1][0])\n xt1 = int((x1 + x2 + x3 + x4 + x5 + x6 + x7 + x8 + x9) / 9)\n xt2 = int(-x1 - (2 * x2) - x3 + x7 + (2 * x8) + x9)\n xt3 = int(-x1 - (2 * x4) - x7 + x3 + (2 * x6) + x9)\n xb = int((2 * xt1) + xt2 + xt3)\n if xb < 0:\n xb = -xb\n if xb > 255:\n xb = 255\n imgSharpnessGrayL[y][x] = (xb, xb, xb)\n \nplt.imshow(imgSharpnessGrayL)\nplt.title(\"Sharpness Gray 2:1\")\nplt.show()", "_____no_output_____" ] ], [ [ "------------------", "_____no_output_____" ], [ "## Sharpness Gray (1:2 LPF:HPF)", "_____no_output_____" ] ], [ [ "imgSharpnessGrayH = np.zeros((imgNormal.shape[0], imgNormal.shape[1], 3), dtype=np.uint8)\n\nfor y in range(1, imgNormal.shape[0] - 1):\n for x in range(1, imgNormal.shape[1] - 1):\n x1 = int(imgGrayscale[y - 1][x - 1][0])\n x2 = int(imgGrayscale[y][x - 1][0])\n x3 = int(imgGrayscale[y + 1][x - 1][0])\n x4 = int(imgGrayscale[y - 1][x][0])\n x5 = int(imgGrayscale[y][x][0])\n x6 = int(imgGrayscale[y + 1][x][0])\n x7 = int(imgGrayscale[y - 1][x + 1][0])\n x8 = int(imgGrayscale[y][x + 1][0])\n x9 = int(imgGrayscale[y + 1][x + 1][0])\n xt1 = int((x1 + x2 + x3 + x4 + x5 + x6 + x7 + x8 + x9) / 9)\n xt2 = int(-x1 - (2 * x2) - x3 + x7 + (2 * x8) + x9)\n xt3 = int(-x1 - (2 * x4) - x7 + x3 + (2 * x6) + x9)\n xb = int(xt1 + (2 * xt2) + (2 * xt3))\n if xb < 0:\n xb = -xb\n if xb > 255:\n xb = 255\n imgSharpnessGrayH[y][x] = (xb, xb, xb)\n \nplt.imshow(imgSharpnessGrayH)\nplt.title(\"Sharpness Gray 1:2\")\nplt.show()", "_____no_output_____" ] ], [ [ "---------------------", "_____no_output_____" ] ] ]

[ "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown", "code", "markdown" ]

[ [ "markdown", "markdown", "markdown", "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown", "markdown" ], [ "code" ], [ "markdown" ] ]

[ "MIT" ]

[ "MIT" ]