python.machine.learning/main.py at master · demyashev/python.machine.learning · GitHub

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
import numpy as np
import nltk
import os
import email_read_util

from sklearn.naive_bayes import MultinomialNB
from sklearn.ensemble import RandomForestClassifier
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import cross_val_score
from nltk.metrics import edit_distance


def read_email_files():
    X = []
    y = []
    for i in range(len(labels)):
        filename = 'inmail.' + str(i+1)
        email_str = email_read_util.extract_email_text(
            os.path.join(DATA_DIR, filename))
        X.append(email_str)
        y.append(labels[filename])
    return X, y


def read_email_files_load():
    X = []
    y = []
    for i in range(len(labels)):
        filename = 'inmail.' + str(i+1)
        email_str = email_read_util.load(
            os.path.join(DATA_DIR, filename))
        X.append(email_str)
        y.append(labels[filename])
    return X, y


def classify(X, y, clf, vectorizer):

    # Преобразование массива строк в структуру bag of words
    X_vector = vectorizer.fit_transform(X)

    # Обучение
    clf.fit(X_vector, y)

    # Оценка
    score = cross_val_score(clf, X_vector, y, cv=3)
    print('Accuracy: ', end='')
    print(score)
    print('Mean accuracy: ', end='')
    print(score.mean())


def compare(email_str0, email_str1, clf, vectorizer):

    # Классификация исходного письма
    Z = []
    Z.append(email_str0)
    Z_vector = vectorizer.transform(Z)
    label = clf.predict(Z_vector)[0]
    print('Predicted label: ', end='')
    print(label)

    # Классификация измененного письма
    Z = []
    Z.append(email_str1)
    Z_vector = vectorizer.transform(Z)
    label = clf.predict(Z_vector)[0]
    print('New label: ', end='')
    print(label)


nltk.download('punkt')
nltk.download('stopwords')

DATA_DIR = 'datasets/trec07p/data/'
LABELS_FILE = 'datasets/trec07p/full/index'

# Получаем метки классов
labels = {}
with open(LABELS_FILE) as f:
    for line in f:
        line = line.strip()
        label, key = line.split()
        labels[key.split('/')[-1]] = 1 if label.lower() == 'ham' else 0


print('Наивный Байес')
X, y = read_email_files()
vectorizer = CountVectorizer()
clf = MultinomialNB()
classify(X, y, clf, vectorizer)

# Сравнение измененного письма (extract_email_text)
print('Отравление Байеса')
filename = 'inmail.4'
email_str0 = email_read_util.extract_email_text(os.path.join(DATA_DIR, filename))
email_str1 = email_read_util.extract_email_text(os.path.join(filename))
ind = X.index(email_str0)
print('First label: ', end='')
print(y[ind])
print('Edit distance: ', end='')
print(edit_distance(email_str0, email_str1))

compare(email_str0, email_str1, clf, vectorizer)

print('Замена extract_email_text на load')
X, y = read_email_files_load()
classify(X, y, clf, vectorizer)
email_str0 = email_read_util.load(os.path.join(DATA_DIR, filename))
email_str1 = email_read_util.load(os.path.join(filename))
compare(email_str0, email_str1, clf, vectorizer)

print('Биграммы')
vectorizer = CountVectorizer(ngram_range=(2, 2))
classify(X, y, clf, vectorizer)
compare(email_str0, email_str1, clf, vectorizer)

print('TF/IDF')
vectorizer = TfidfVectorizer()
classify(X, y, clf, vectorizer)
compare(email_str0, email_str1, clf, vectorizer)

print('Случайный лес')
clf = RandomForestClassifier()
classify(X, y, clf, vectorizer)
compare(email_str0, email_str1, clf, vectorizer)