File size: 10,418 Bytes
0d85011
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
import numpy as np
import pandas as pd
import nltk, string, logging, pickle, torch
import matplotlib.pyplot as plt
import seaborn as sns
from collections import Counter
from nltk.corpus import stopwords
from sklearn.metrics import confusion_matrix, classification_report
from sklearn.model_selection import train_test_split, cross_val_score
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.svm import SVC
from sklearn.naive_bayes import MultinomialNB
from sklearn.ensemble import ExtraTreesClassifier, VotingClassifier
from sklearn.metrics import accuracy_score, precision_score, f1_score
from torch.cuda import is_available as cuda_available

# Setup logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

class MonarchButterflyOptimizer:
    def __init__(self, bounds, n_butterflies=20, p_period=1.2, migration_ratio=0.85, max_iter=30, use_gpu=False):
        self.bounds = bounds
        self.n_butterflies = n_butterflies
        self.p_period = p_period
        self.migration_ratio = migration_ratio
        self.max_iter = max_iter
        self.best_solution = None
        self.best_fitness = float('-inf')
        
        # GPU setup
        self.use_gpu = use_gpu and cuda_available()
        self.device = torch.device('cuda' if self.use_gpu else 'cpu')
        logger.info(f"Using device: {self.device}")

    def initialize(self):
        try:
            population = []
            for _ in range(self.n_butterflies):
                butterfly = {}
                for param, (low, high) in self.bounds.items():
                    if isinstance(low, int) and isinstance(high, int):
                        butterfly[param] = int(torch.randint(low, high+1, (1,), device=self.device).item())
                    else:
                        butterfly[param] = float(torch.rand(1, device=self.device).item() * (high - low) + low)
                population.append(butterfly)
            return population
        except RuntimeError as e:
            logger.error(f"CUDA error during initialization: {e}")
            self.device = torch.device('cpu')
            logger.info("Falling back to CPU")
            return self.initialize()

    def migration(self, population):
        try:
            new_population = []
            migration_tensor = torch.rand(len(population), device=self.device)
            
            for idx, butterfly in enumerate(population):
                if migration_tensor[idx].item() < self.migration_ratio:
                    new_butterfly = {}
                    for param in butterfly:
                        r = torch.rand(1, device=self.device).item()
                        new_val = butterfly[param] + self.p_period * r * (self.best_solution[param] - butterfly[param])
                        new_butterfly[param] = self.clip(new_val, param)
                    new_population.append(new_butterfly)
                else:
                    new_population.append(butterfly.copy())
            return new_population
        except RuntimeError as e:
            logger.error(f"CUDA error during migration: {e}")
            self.device = torch.device('cpu')
            logger.info("Falling back to CPU")
            return self.migration(population)
    
    def clip(self, value, param):
        low, high = self.bounds[param]
        if isinstance(low, int) and isinstance(high, int):
            return int(np.clip(value, low, high))
        return np.clip(value, low, high)
    
    def optimize(self, fitness_func):
        population = self.initialize()
        
        for _ in range(self.max_iter):
            for butterfly in population:
                fitness = fitness_func(butterfly)
                if fitness > self.best_fitness:
                    self.best_fitness = fitness
                    self.best_solution = butterfly.copy()
            
            population = self.migration(population)
        
        return self.best_solution, self.best_fitness
    
def plot_dataset_insights(df):
    plt.figure(figsize=(15, 5))
    
    plt.subplot(131)
    sns.histplot(data=df, x='feature_length', hue='target', bins=50)
    plt.title('Message Length Distribution')
    
    plt.subplot(132)
    df['target'].value_counts().plot(kind='bar')
    plt.title('Class Distribution')
    
    plt.subplot(133)
    sns.boxplot(data=df, x='target', y='word_count')
    plt.title('Word Count by Class')
    
    plt.tight_layout()
    plt.savefig('./graphs/dataset_insights.png')
    plt.close()

def plot_word_clouds(df):
    from wordcloud import WordCloud
    plt.figure(figsize=(15, 5))
    
    for idx, label in enumerate(['ham', 'spam']):
        text = ' '.join(df[df['target'] == label]['transformed_text'])
        wordcloud = WordCloud(width=800, height=400).generate(text)
        
        plt.subplot(1, 2, idx+1)
        plt.imshow(wordcloud)
        plt.axis('off')
        plt.title(f'Word Cloud - {label.upper()}')
    
    plt.savefig('./graphs/wordclouds.png')
    plt.close()

def plot_performance_metrics(y_test, y_pred, model):
    plt.figure(figsize=(15, 5))
    
    plt.subplot(131)
    cm = confusion_matrix(y_test, y_pred)
    sns.heatmap(cm, annot=True, fmt='d', cmap='Blues')
    plt.title('Confusion Matrix')
    
    plt.subplot(132)
    report = classification_report(y_test, y_pred, output_dict=True)
    sns.heatmap(pd.DataFrame(report).iloc[:-1, :].T, annot=True, cmap='RdYlGn')
    plt.title('Classification Report')
    
    plt.subplot(133)
    etc = model.named_estimators_['etc']
    importances = pd.Series(etc.feature_importances_)
    importances.nlargest(10).plot(kind='bar')
    plt.title('Top 10 Important Features')
    
    plt.tight_layout()
    plt.savefig('./graphs/performance_metrics.png')
    plt.close()

def save_metrics(metrics):
    with open('./models/metrics.txt', 'w') as f:
        for metric, value in metrics.items():
            f.write(f"{metric}: {value:.4f}\n")

def create_optimized_ensemble(X_train, y_train, mbo_params):
    param_bounds = {
        'svc_C': (0.1, 20.0),
        'svc_gamma': (0.001, 1.0),
        'mnb_alpha': (0.1, 2.0),
        'etc_n_estimators': (100, 300),
        'w1': (0, 5),
        'w2': (0, 5),
        'w3': (0, 5)
    }
    
    mbo = MonarchButterflyOptimizer(
        param_bounds,
        n_butterflies=int(mbo_params.get('n_butterflies', 20)),
        p_period=float(mbo_params.get('p_period', 1.2)),
        migration_ratio=float(mbo_params.get('migration_ratio', 0.85)),
        max_iter=int(mbo_params.get('max_iter', 30)),
        use_gpu=bool(mbo_params.get('use_gpu', False))
    )
    
    def fitness_function(params):
        svc = SVC(kernel='rbf', C=params['svc_C'], 
                  gamma=params['svc_gamma'], probability=True)
        mnb = MultinomialNB(alpha=params['mnb_alpha'])
        etc = ExtraTreesClassifier(n_estimators=int(params['etc_n_estimators']))
        
        estimators = [('svc', svc), ('mnb', mnb), ('etc', etc)]
        weights = [params['w1'], params['w2'], params['w3']]
        
        clf = VotingClassifier(estimators=estimators, voting='soft', weights=weights)
        scores = cross_val_score(clf, X_train, y_train, cv=5)
        return np.mean(scores)
    
    # Initialize and run MBO
    mbo = MonarchButterflyOptimizer(param_bounds)
    best_params, _ = mbo.optimize(fitness_function)
    # Create final model with optimized parameters
    svc = SVC(kernel='rbf', C=best_params['svc_C'], 
              gamma=best_params['svc_gamma'], probability=True)
    mnb = MultinomialNB(alpha=best_params['mnb_alpha'])
    etc = ExtraTreesClassifier(n_estimators=int(best_params['etc_n_estimators']))
    
    estimators = [('svc', svc), ('mnb', mnb), ('etc', etc)]
    weights = [best_params['w1'], best_params['w2'], best_params['w3']]
    
    return VotingClassifier(estimators=estimators, voting='soft', weights=weights)

def main(mbo_params=None):
    try:
        logger.info("Loading data...")
        # Load and preprocess data
        df = pd.read_csv('./data/spam.csv', encoding='latin-1')
        df = df.drop(['Unnamed: 2', 'Unnamed: 3', 'Unnamed: 4'], axis=1)
        df = df.rename(columns={'v1': 'target', 'v2': 'text'})
        
        logger.info("Preprocessing text...")
        df['transformed_text'] = df['text'].apply(lambda x: x.lower().translate(str.maketrans('', '', string.punctuation)))
        df['word_count'] = df['transformed_text'].str.split().str.len()
        df['feature_length'] = df['transformed_text'].apply(len)
        
        logger.info("Generating visualizations...")
        plot_dataset_insights(df)
        plot_word_clouds(df)
        
        tfidf = TfidfVectorizer(max_features=5000, ngram_range=(1,3))
        X = tfidf.fit_transform(df['transformed_text'])
        y = (df['target'] == 'spam').astype(int)
        
        X_train, X_test, y_train, y_test = train_test_split(
            X, y, test_size=0.2, random_state=42, stratify=y
        )
        
        logger.info("Training model with MBO...")
        if mbo_params and mbo_params.get('use_gpu'):
            logger.info("GPU acceleration enabled")
        model = create_optimized_ensemble(X_train, y_train, mbo_params or {})
        
        model.fit(X_train, y_train)
        
        y_pred = model.predict(X_test)
        
        metrics = {
            "Accuracy": accuracy_score(y_test, y_pred),
            "Precision": precision_score(y_test, y_pred),
            "F1": f1_score(y_test, y_pred)
        }
        
        save_metrics(metrics)
        for metric, value in metrics.items():
            logger.info(f"{metric}: {value:.4f}")
            
        plot_performance_metrics(y_test, y_pred, model)
        
        logger.info("Saving models...")
        with open('./models/vectorizer_mbo.pkl', 'wb') as f:
            pickle.dump(tfidf, f)
        with open('./models/model_mbo.pkl', 'wb') as f:
            pickle.dump(model, f)
            
        logger.info("MBO optimization completed successfully")
        
    except Exception as e:
        logger.error(f"An error occurred: {e}")
        raise

if __name__ == "__main__":
    main()