Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
@@ -1,122 +1,19 @@
|
|
1 |
---
|
2 |
-
|
3 |
-
|
4 |
-
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
-
|
15 |
-
|
16 |
-
split: validation
|
17 |
-
- path: ag_news/test.csv
|
18 |
-
split: test
|
19 |
-
- config_name: amazon_polarity
|
20 |
-
data_files:
|
21 |
-
- path: amazon_polarity/train.csv
|
22 |
-
split: train
|
23 |
-
- path: amazon_polarity/validation.csv
|
24 |
-
split: validation
|
25 |
-
- path: amazon_polarity/test.csv
|
26 |
-
split: test
|
27 |
-
- config_name: banking77
|
28 |
-
data_files:
|
29 |
-
- split: train
|
30 |
-
path: banking77/train-*
|
31 |
-
- split: test
|
32 |
-
path: banking77/test-*
|
33 |
-
- split: validation
|
34 |
-
path: banking77/validation-*
|
35 |
-
- config_name: emotion
|
36 |
-
data_files:
|
37 |
-
- path: emotion/train.csv
|
38 |
-
split: train
|
39 |
-
- path: emotion/validation.csv
|
40 |
-
split: validation
|
41 |
-
- path: emotion/test.csv
|
42 |
-
split: test
|
43 |
-
- config_name: imdb
|
44 |
-
data_files:
|
45 |
-
- path: imdb/train.csv
|
46 |
-
split: train
|
47 |
-
- path: imdb/validation.csv
|
48 |
-
split: validation
|
49 |
-
- path: imdb/test.csv
|
50 |
-
split: test
|
51 |
-
- config_name: trec
|
52 |
-
data_files:
|
53 |
-
- split: train
|
54 |
-
path: trec/train-*
|
55 |
-
- split: test
|
56 |
-
path: trec/test-*
|
57 |
-
- split: validation
|
58 |
-
path: trec/validation-*
|
59 |
-
- config_name: twenty_newsgroups
|
60 |
-
data_files:
|
61 |
-
- path: twenty_newsgroups/train.csv
|
62 |
-
split: train
|
63 |
-
- path: twenty_newsgroups/validation.csv
|
64 |
-
split: validation
|
65 |
-
- path: twenty_newsgroups/test.csv
|
66 |
-
split: test
|
67 |
-
- config_name: yelp_polarity
|
68 |
-
data_files:
|
69 |
-
- path: yelp_polarity/train.csv
|
70 |
-
split: train
|
71 |
-
- path: yelp_polarity/validation.csv
|
72 |
-
split: validation
|
73 |
-
- path: yelp_polarity/test.csv
|
74 |
-
split: test
|
75 |
dataset_info:
|
76 |
-
- config_name: MASSIVE
|
77 |
-
features:
|
78 |
-
- name: text
|
79 |
-
dtype: string
|
80 |
-
- name: label
|
81 |
-
dtype: int64
|
82 |
-
splits:
|
83 |
-
- name: train
|
84 |
-
num_bytes: 541604
|
85 |
-
num_examples: 11514
|
86 |
-
- name: validation
|
87 |
-
num_bytes: 95125
|
88 |
-
num_examples: 2033
|
89 |
-
- name: test
|
90 |
-
num_bytes: 138473
|
91 |
-
num_examples: 2974
|
92 |
-
download_size: 379162
|
93 |
-
dataset_size: 775202
|
94 |
-
- config_name: ag_news
|
95 |
-
features:
|
96 |
-
- dtype: string
|
97 |
-
name: text
|
98 |
-
- dtype: int64
|
99 |
-
name: label
|
100 |
-
splits:
|
101 |
-
- name: train
|
102 |
-
num_examples: 90000
|
103 |
-
- name: validation
|
104 |
-
num_examples: 30000
|
105 |
-
- name: test
|
106 |
-
num_examples: 7600
|
107 |
-
- config_name: amazon_polarity
|
108 |
-
features:
|
109 |
-
- dtype: string
|
110 |
-
name: text
|
111 |
-
- dtype: int64
|
112 |
-
name: label
|
113 |
-
splits:
|
114 |
-
- name: train
|
115 |
-
num_examples: 2700000
|
116 |
-
- name: validation
|
117 |
-
num_examples: 900000
|
118 |
-
- name: test
|
119 |
-
num_examples: 400000
|
120 |
- config_name: banking77
|
121 |
features:
|
122 |
- name: text
|
@@ -125,42 +22,11 @@ dataset_info:
|
|
125 |
dtype: int64
|
126 |
splits:
|
127 |
- name: train
|
128 |
-
num_bytes: 535404
|
129 |
num_examples: 7502
|
130 |
- name: test
|
131 |
-
num_bytes: 204010
|
132 |
num_examples: 3080
|
133 |
- name: validation
|
134 |
-
num_bytes: 179624
|
135 |
num_examples: 2501
|
136 |
-
download_size: 452800
|
137 |
-
dataset_size: 919038
|
138 |
-
- config_name: emotion
|
139 |
-
features:
|
140 |
-
- dtype: string
|
141 |
-
name: text
|
142 |
-
- dtype: int64
|
143 |
-
name: label
|
144 |
-
splits:
|
145 |
-
- name: train
|
146 |
-
num_examples: 250085
|
147 |
-
- name: validation
|
148 |
-
num_examples: 83362
|
149 |
-
- name: test
|
150 |
-
num_examples: 41681
|
151 |
-
- config_name: imdb
|
152 |
-
features:
|
153 |
-
- dtype: string
|
154 |
-
name: text
|
155 |
-
- dtype: int64
|
156 |
-
name: label
|
157 |
-
splits:
|
158 |
-
- name: train
|
159 |
-
num_examples: 18750
|
160 |
-
- name: validation
|
161 |
-
num_examples: 6250
|
162 |
-
- name: test
|
163 |
-
num_examples: 25000
|
164 |
- config_name: trec
|
165 |
features:
|
166 |
- name: text
|
@@ -169,265 +35,118 @@ dataset_info:
|
|
169 |
dtype: int64
|
170 |
splits:
|
171 |
- name: train
|
172 |
-
num_bytes: 255609
|
173 |
num_examples: 4089
|
174 |
- name: test
|
175 |
-
num_bytes: 23979
|
176 |
num_examples: 500
|
177 |
- name: validation
|
178 |
-
num_bytes: 85861
|
179 |
num_examples: 1363
|
180 |
-
|
181 |
-
dataset_size: 365449
|
182 |
-
- config_name: twenty_newsgroups
|
183 |
features:
|
184 |
-
-
|
185 |
-
|
186 |
-
-
|
187 |
-
|
188 |
splits:
|
189 |
- name: train
|
190 |
-
num_examples:
|
191 |
-
- name: validation
|
192 |
-
num_examples: 2829
|
193 |
- name: test
|
194 |
-
num_examples:
|
195 |
-
-
|
|
|
|
|
196 |
features:
|
197 |
-
-
|
198 |
-
|
199 |
-
-
|
200 |
-
|
201 |
splits:
|
202 |
- name: train
|
203 |
-
num_examples:
|
204 |
-
- name: validation
|
205 |
-
num_examples: 140000
|
206 |
- name: test
|
207 |
-
num_examples:
|
208 |
-
|
209 |
-
|
210 |
-
|
211 |
-
|
212 |
-
|
213 |
-
-
|
214 |
-
|
215 |
-
-
|
216 |
-
-
|
217 |
-
-
|
218 |
-
-
|
219 |
-
|
220 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
221 |
---
|
222 |
-
|
223 |
# Data Preprocessing AutoML Benchmarks
|
224 |
|
225 |
This repository contains text classification datasets with known data quality issues for preprocessing research in AutoML.
|
226 |
|
227 |
-
## Dataset Categories
|
228 |
-
|
229 |
-
### Redundancy Issues
|
230 |
-
- **ag_news**: News categorization with topic overlap
|
231 |
-
- **twenty_newsgroups**: Newsgroup posts with cross-posting
|
232 |
-
|
233 |
-
### Class Imbalance Issues
|
234 |
-
- **yelp_polarity**: Sentiment analysis with rating bias
|
235 |
-
|
236 |
-
### Label Noise Issues
|
237 |
-
- **imdb**: Movie reviews with subjective labels
|
238 |
-
- **amazon_polarity**: Product reviews with rating inconsistencies
|
239 |
-
|
240 |
-
### Outlier Issues
|
241 |
-
- **emotion**: Twitter emotion with length outliers
|
242 |
-
|
243 |
-
## Dataset Structure
|
244 |
-
|
245 |
-
Each dataset contains:
|
246 |
-
- `train.csv`: Training split (~75% of original training data)
|
247 |
-
- `validation.csv`: Validation split (~25% of original training data)
|
248 |
-
- `test.csv`: Test split (original test set preserved)
|
249 |
-
|
250 |
-
All datasets have consistent columns:
|
251 |
-
- `text`: Input text
|
252 |
-
- `label`: Target label (integer encoded)
|
253 |
-
|
254 |
-
**Important**: Original test sets are preserved to maintain methodological integrity and enable comparison with published benchmarks.
|
255 |
-
|
256 |
## Usage
|
257 |
|
|
|
|
|
258 |
```python
|
259 |
from datasets import load_dataset
|
260 |
|
261 |
-
#
|
262 |
-
dataset = load_dataset("MothMalone/data-preprocessing-automl-benchmarks", "
|
263 |
-
|
264 |
-
# Access splits
|
265 |
-
train_data = dataset["train"]
|
266 |
-
val_data = dataset["validation"]
|
267 |
-
test_data = dataset["test"]
|
268 |
```
|
269 |
|
270 |
-
##
|
271 |
-
|
272 |
-
|
273 |
-
|
274 |
-
|
275 |
-
|
276 |
-
|
277 |
-
|
278 |
-
|
279 |
-
|
280 |
-
|
281 |
-
|
282 |
-
|
283 |
-
|
284 |
-
|
285 |
-
|
286 |
-
|
287 |
-
|
288 |
-
|
289 |
-
|
290 |
-
|
291 |
-
|
292 |
-
|
293 |
-
|
294 |
-
|
295 |
-
|
296 |
-
|
297 |
-
|
298 |
-
|
299 |
-
|
300 |
-
|
301 |
-
|
302 |
-
|
303 |
-
|
304 |
-
|
305 |
-
quality_issues:
|
306 |
-
- label_noise
|
307 |
-
- rating_inconsistency
|
308 |
-
target_column: label
|
309 |
-
task_type: binary_classification
|
310 |
-
test_samples: 400000
|
311 |
-
text_columns:
|
312 |
-
- text
|
313 |
-
total_samples: 4000000
|
314 |
-
train_samples: 2700000
|
315 |
-
validation_samples: 900000
|
316 |
-
emotion:
|
317 |
-
class_names:
|
318 |
-
- sadness
|
319 |
-
- joy
|
320 |
-
- love
|
321 |
-
- anger
|
322 |
-
- fear
|
323 |
-
- surprise
|
324 |
-
description: Twitter emotion classification with text length outliers
|
325 |
-
name: Emotion Classification
|
326 |
-
num_classes: 6
|
327 |
-
original_test_samples: 41681
|
328 |
-
original_train_samples: 333447
|
329 |
-
quality_issues:
|
330 |
-
- length_outliers
|
331 |
-
- text_anomalies
|
332 |
-
target_column: label
|
333 |
-
task_type: multi_classification
|
334 |
-
test_samples: 41681
|
335 |
-
text_columns:
|
336 |
-
- text
|
337 |
-
total_samples: 375128
|
338 |
-
train_samples: 250085
|
339 |
-
validation_samples: 83362
|
340 |
-
imdb:
|
341 |
-
class_names:
|
342 |
-
- negative
|
343 |
-
- positive
|
344 |
-
description: Movie reviews with subjective sentiment labels and borderline cases
|
345 |
-
name: IMDB Movie Reviews
|
346 |
-
num_classes: 2
|
347 |
-
original_test_samples: 25000
|
348 |
-
original_train_samples: 25000
|
349 |
-
quality_issues:
|
350 |
-
- label_noise
|
351 |
-
- subjective_labels
|
352 |
-
- borderline_cases
|
353 |
-
target_column: label
|
354 |
-
task_type: binary_classification
|
355 |
-
test_samples: 25000
|
356 |
-
text_columns:
|
357 |
-
- text
|
358 |
-
total_samples: 50000
|
359 |
-
train_samples: 18750
|
360 |
-
validation_samples: 6250
|
361 |
-
twenty_newsgroups:
|
362 |
-
class_names:
|
363 |
-
- alt.atheism
|
364 |
-
- comp.graphics
|
365 |
-
- comp.os.ms-windows.misc
|
366 |
-
- comp.sys.ibm.pc.hardware
|
367 |
-
- comp.sys.mac.hardware
|
368 |
-
- comp.windows.x
|
369 |
-
- misc.forsale
|
370 |
-
- rec.autos
|
371 |
-
- rec.motorcycles
|
372 |
-
- rec.sport.baseball
|
373 |
-
- rec.sport.hockey
|
374 |
-
- sci.crypt
|
375 |
-
- sci.electronics
|
376 |
-
- sci.med
|
377 |
-
- sci.space
|
378 |
-
- soc.religion.christian
|
379 |
-
- talk.politics.guns
|
380 |
-
- talk.politics.mideast
|
381 |
-
- talk.politics.misc
|
382 |
-
- talk.religion.misc
|
383 |
-
description: Newsgroup posts with overlapping topics and cross-posting
|
384 |
-
name: 20 Newsgroups
|
385 |
-
num_classes: 20
|
386 |
-
original_test_samples: 7532
|
387 |
-
original_train_samples: 11314
|
388 |
-
quality_issues:
|
389 |
-
- redundancy
|
390 |
-
- cross_posting
|
391 |
-
- similar_topics
|
392 |
-
target_column: label
|
393 |
-
task_type: multi_classification
|
394 |
-
test_samples: 7532
|
395 |
-
text_columns:
|
396 |
-
- text
|
397 |
-
total_samples: 18846
|
398 |
-
train_samples: 8485
|
399 |
-
validation_samples: 2829
|
400 |
-
yelp_polarity:
|
401 |
-
class_names:
|
402 |
-
- negative
|
403 |
-
- positive
|
404 |
-
description: Yelp reviews with positive/negative sentiment, naturally imbalanced
|
405 |
-
name: Yelp Review Polarity
|
406 |
-
num_classes: 2
|
407 |
-
original_test_samples: 38000
|
408 |
-
original_train_samples: 560000
|
409 |
-
quality_issues:
|
410 |
-
- moderate_imbalance
|
411 |
-
- rating_bias
|
412 |
-
target_column: label
|
413 |
-
task_type: binary_classification
|
414 |
-
test_samples: 38000
|
415 |
-
text_columns:
|
416 |
-
- text
|
417 |
-
total_samples: 598000
|
418 |
-
train_samples: 420000
|
419 |
-
validation_samples: 140000
|
420 |
-
|
421 |
-
|
422 |
-
## Citation
|
423 |
-
|
424 |
-
If you use these datasets in your research, please cite the original sources and this collection:
|
425 |
-
|
426 |
-
```bibtex
|
427 |
-
@misc{mothmalone2024preprocessing,
|
428 |
-
title={Data Preprocessing AutoML Benchmarks},
|
429 |
-
author={MothMalone},
|
430 |
-
year={2024},
|
431 |
-
url={https://huggingface.co/datasets/MothMalone/data-preprocessing-automl-benchmarks}
|
432 |
-
}
|
433 |
-
```
|
|
|
1 |
---
|
2 |
+
license: apache-2.0
|
3 |
+
task_categories:
|
4 |
+
- text-classification
|
5 |
+
language:
|
6 |
+
- en
|
7 |
+
tags:
|
8 |
+
- data-preprocessing
|
9 |
+
- automl
|
10 |
+
- benchmarks
|
11 |
+
size_categories:
|
12 |
+
- n<1K
|
13 |
+
- 1K<n<10K
|
14 |
+
- 10K<n<100K
|
15 |
+
- 100K<n<1M
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
16 |
dataset_info:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
17 |
- config_name: banking77
|
18 |
features:
|
19 |
- name: text
|
|
|
22 |
dtype: int64
|
23 |
splits:
|
24 |
- name: train
|
|
|
25 |
num_examples: 7502
|
26 |
- name: test
|
|
|
27 |
num_examples: 3080
|
28 |
- name: validation
|
|
|
29 |
num_examples: 2501
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
- config_name: trec
|
31 |
features:
|
32 |
- name: text
|
|
|
35 |
dtype: int64
|
36 |
splits:
|
37 |
- name: train
|
|
|
38 |
num_examples: 4089
|
39 |
- name: test
|
|
|
40 |
num_examples: 500
|
41 |
- name: validation
|
|
|
42 |
num_examples: 1363
|
43 |
+
- config_name: financial_phrasebank
|
|
|
|
|
44 |
features:
|
45 |
+
- name: text
|
46 |
+
dtype: string
|
47 |
+
- name: label
|
48 |
+
dtype: int64
|
49 |
splits:
|
50 |
- name: train
|
51 |
+
num_examples: 1698
|
|
|
|
|
52 |
- name: test
|
53 |
+
num_examples: 0
|
54 |
+
- name: validation
|
55 |
+
num_examples: 566
|
56 |
+
- config_name: MASSIVE
|
57 |
features:
|
58 |
+
- name: text
|
59 |
+
dtype: string
|
60 |
+
- name: label
|
61 |
+
dtype: int64
|
62 |
splits:
|
63 |
- name: train
|
64 |
+
num_examples: 11514
|
|
|
|
|
65 |
- name: test
|
66 |
+
num_examples: 2974
|
67 |
+
- name: validation
|
68 |
+
num_examples: 2033
|
69 |
+
configs:
|
70 |
+
- config_name: banking77
|
71 |
+
data_files:
|
72 |
+
- split: train
|
73 |
+
path: data/banking77-train.parquet
|
74 |
+
- split: test
|
75 |
+
path: data/banking77-test.parquet
|
76 |
+
- split: validation
|
77 |
+
path: data/banking77-validation.parquet
|
78 |
+
- config_name: trec
|
79 |
+
data_files:
|
80 |
+
- split: train
|
81 |
+
path: data/trec-train.parquet
|
82 |
+
- split: test
|
83 |
+
path: data/trec-test.parquet
|
84 |
+
- split: validation
|
85 |
+
path: data/trec-validation.parquet
|
86 |
+
- config_name: financial_phrasebank
|
87 |
+
data_files:
|
88 |
+
- split: train
|
89 |
+
path: data/financial_phrasebank-train.parquet
|
90 |
+
- split: test
|
91 |
+
path: data/financial_phrasebank-test.parquet
|
92 |
+
- split: validation
|
93 |
+
path: data/financial_phrasebank-validation.parquet
|
94 |
+
- config_name: MASSIVE
|
95 |
+
data_files:
|
96 |
+
- split: train
|
97 |
+
path: data/MASSIVE-train.parquet
|
98 |
+
- split: test
|
99 |
+
path: data/MASSIVE-test.parquet
|
100 |
+
- split: validation
|
101 |
+
path: data/MASSIVE-validation.parquet
|
102 |
---
|
|
|
103 |
# Data Preprocessing AutoML Benchmarks
|
104 |
|
105 |
This repository contains text classification datasets with known data quality issues for preprocessing research in AutoML.
|
106 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
107 |
## Usage
|
108 |
|
109 |
+
Load a specific dataset configuration like this:
|
110 |
+
|
111 |
```python
|
112 |
from datasets import load_dataset
|
113 |
|
114 |
+
# Example for loading the TREC dataset
|
115 |
+
dataset = load_dataset("MothMalone/data-preprocessing-automl-benchmarks", "trec")
|
|
|
|
|
|
|
|
|
|
|
116 |
```
|
117 |
|
118 |
+
## Available Datasets
|
119 |
+
|
120 |
+
Below are the details for each dataset configuration available in this repository.
|
121 |
+
|
122 |
+
### banking77
|
123 |
+
- Description:
|
124 |
+
- Data Quality Issue: N/A
|
125 |
+
- Classes: 77
|
126 |
+
- Training Samples: 7502
|
127 |
+
- Validation Samples: 2501
|
128 |
+
- Test Samples: 3080
|
129 |
+
|
130 |
+
### trec
|
131 |
+
- Description: The Text REtrieval Conference (TREC) Question Classification dataset contains 5500 labeled questions in training set and another 500 for test set.
|
132 |
+
- Data Quality Issue: N/A
|
133 |
+
- Classes: 6
|
134 |
+
- Training Samples: 4089
|
135 |
+
- Validation Samples: 1363
|
136 |
+
- Test Samples: 500
|
137 |
+
|
138 |
+
### financial_phrasebank
|
139 |
+
- Description: The key arguments for the low utilization of statistical techniques in
|
140 |
+
- Data Quality Issue: N/A
|
141 |
+
- Classes: 3
|
142 |
+
- Training Samples: 1698
|
143 |
+
- Validation Samples: 566
|
144 |
+
- Test Samples: 0
|
145 |
+
|
146 |
+
### MASSIVE
|
147 |
+
- Description:
|
148 |
+
- Data Quality Issue: N/A
|
149 |
+
- Classes: 60
|
150 |
+
- Training Samples: 11514
|
151 |
+
- Validation Samples: 2033
|
152 |
+
- Test Samples: 2974
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|