mp1/task1.py at main · shvnks/mp1 · 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
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
# Authors: Anthony Wong, Derek Lam, Tyler Shanks
# Due: October 18, 2021
# COMP-472

# Importing necessary libraries
import numpy as np
from sklearn.datasets import load_files
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import train_test_split
from sklearn.naive_bayes import MultinomialNB
from sklearn.metrics import accuracy_score, f1_score, classification_report, confusion_matrix, plot_confusion_matrix
from matplotlib import pyplot as plt
import os

# Step 2
# 510 files
businessList = os.listdir('BBC/business')
business = len(businessList)
# 386 files
entertainmentList = os.listdir('BBC/entertainment')
entertainment = len(entertainmentList)
# 417 files
politicList = os.listdir('BBC/politics')
politics = len(politicList)
# 511 files
sportList = os.listdir('BBC/sport')
sport = len(sportList)
# 401 files
techList = os.listdir('BBC/tech')
tech = len(techList)

# plotting chart
dev_x = ['business', 'entertainment', 'politics', 'sport', 'tech']
dev_y = [business, entertainment, politics, sport, tech]

plt.bar(dev_x, dev_y)
plt.title("Documents per section")
plt.ylabel("Number of documents")
plt.tight_layout()

plt.savefig("output/task1/bbc-distribution.pdf", bbox_inches="tight")

# Step 3
files = load_files('BBC', encoding='latin1')

# Step 4
labels, counts = np.unique(files.target, return_counts=True)
labels_str = np.array(files.target_names)[labels]
print(dict(zip(labels_str, counts)))

# Step 5
X_train, X_test, y_train, y_test = train_test_split(files.data, files.target, train_size=0.8, test_size=0.2)
print("training set:", len(X_train), "  test set:", len(X_test))

vectorizer = TfidfVectorizer(stop_words="english", decode_error="ignore")  # max_features can be changed
vectorizer.fit(X_train)

tfidf_matrix = vectorizer.transform(X_train)
feature_names = vectorizer.get_feature_names_out()

print(tfidf_matrix.shape)

# Step 6
multiBayes = MultinomialNB()
multiBayes.fit(vectorizer.transform(X_train), y_train)
y_predict = multiBayes.predict(vectorizer.transform(X_test))

# for g), i) and j)
word_per_class = [0, 0, 0, 0, 0]
zeroFreq = [0, 0, 0, 0, 0]
oneFreq = 0
for category in range(0, len(files.target_names)):
    feature = multiBayes.feature_count_[category]
    for count in feature:
        word_per_class[category] += count
        if count == 0:
            zeroFreq[category] += 1
        if count == 1:
            oneFreq += 1

# for k)
my_words = ['Christmas', 'tree']
my_word_matrix = vectorizer.transform(my_words)

# Step 7
fh = open('bbc-performance.txt', 'w')
fh.write("\n********** MultinomialNB default values, try 1 **********\n")
fh.write(str(dict(zip(labels_str, counts))))
fh.write("\nb) Confusion Matrix:\n" + str(confusion_matrix(y_test, y_predict)))
fh.write("\n\nc) Classification Report: \n" + str(classification_report(y_test, y_predict, target_names=labels_str)))
fh.write("\nd) Accuracy:" + str(accuracy_score(y_test, y_predict)))
fh.write("\n   Macro F1:" + str(f1_score(y_test, y_predict, average='macro')))
fh.write("\n   Weighted F1:" + str(f1_score(y_test, y_predict, average='weighted')))
fh.write("\ne) Prior probabilities:" + str(multiBayes.class_log_prior_))
fh.write("\nf) Vocabulary Size:" + str(len(vectorizer.vocabulary_)))
fh.write("\ng) Number of word-tokens in each class:" + str(word_per_class))  # multiBayes.class_count_
fh.write("\nh) Number of word-tokens in entire corpus:" + str(tfidf_matrix.sum()))
fh.write("\ni) Number and percentage of words with zero frequency in each class:" + str(zeroFreq))
fh.write("\nj) Number and percentage of word with one frequency in entire corpus:" + str(oneFreq))
fh.write("\nk) Favorite words: Christmas and tree - " + str(multiBayes.predict_log_proba(my_word_matrix)))


# Step 8 ------------------------------------------------------------------------------------------------------------
multiBayes = MultinomialNB()
multiBayes.fit(vectorizer.transform(X_train), y_train)
y_predict = multiBayes.predict(vectorizer.transform(X_test))

word_per_class = [0, 0, 0, 0, 0]
zeroFreq = [0, 0, 0, 0, 0]
oneFreq = 0
for category in range(0, len(files.target_names)):
    feature = multiBayes.feature_count_[category]
    for count in feature:
        word_per_class[category] += count
        if count == 0:
            zeroFreq[category] += 1
        if count == 1:
            oneFreq += 1

fh.write("\n\n********** MultinomialNB default values, try 2 **********\n")
fh.write(str(dict(zip(labels_str, counts))))
fh.write("\nb) Confusion Matrix:\n" + str(confusion_matrix(y_test, y_predict)))
fh.write("\n\nc) Classification Report: \n" + str(classification_report(y_test, y_predict, target_names=labels_str)))
fh.write("\nd) Accuracy:" + str(accuracy_score(y_test, y_predict)))
fh.write("\n   Macro F1:" + str(f1_score(y_test, y_predict, average='macro')))
fh.write("\n   Weighted F1:" + str(f1_score(y_test, y_predict, average='weighted')))
fh.write("\ne) Prior probabilities:" + str(multiBayes.class_log_prior_))  # not sure
fh.write("\nf) Vocabulary Size:" + str(len(vectorizer.vocabulary_)))
fh.write("\ng) Number of word-tokens in each class:" + str(word_per_class))  # multiBayes.class_count_
fh.write("\nh) Number of word-tokens in entire corpus:" + str(tfidf_matrix.sum()))
fh.write("\ni) Number and percentage of words with zero frequency in each class:" + str(zeroFreq))
fh.write("\nj) Number and percentage of word with one frequency in entire corpus:" + str(oneFreq))
fh.write("\nk) Favorite words: Christmas and tree - " + str(multiBayes.predict_log_proba(my_word_matrix)))


# Step 9 ------------------------------------------------------------------------------------------------------------
multiBayes = MultinomialNB(alpha=0.0001)
multiBayes.fit(vectorizer.transform(X_train), y_train)
y_predict = multiBayes.predict(vectorizer.transform(X_test))

word_per_class = [0, 0, 0, 0, 0]
zeroFreq = [0, 0, 0, 0, 0]
oneFreq = 0
for category in range(0, len(files.target_names)):
    feature = multiBayes.feature_count_[category]
    for count in feature:
        word_per_class[category] += count
        if count == 0:
            zeroFreq[category] += 1
        if count == 1:
            oneFreq += 1

fh.write("\n\n********** MultinomialNB smoothing 0.0001 **********\n")
fh.write(str(dict(zip(labels_str, counts))))
fh.write("\nb) Confusion Matrix:\n" + str(confusion_matrix(y_test, y_predict)))
fh.write("\n\nc) Classification Report: \n" + str(classification_report(y_test, y_predict, target_names=labels_str)))
fh.write("\nd) Accuracy:" + str(accuracy_score(y_test, y_predict)))
fh.write("\n   Macro F1:" + str(f1_score(y_test, y_predict, average='macro')))
fh.write("\n   Weighted F1:" + str(f1_score(y_test, y_predict, average='weighted')))
fh.write("\ne) Prior probabilities:" + str(multiBayes.class_log_prior_))  # not sure
fh.write("\nf) Vocabulary Size:" + str(len(vectorizer.vocabulary_)))
fh.write("\ng) Number of word-tokens in each class:" + str(word_per_class))  # multiBayes.class_count_
fh.write("\nh) Number of word-tokens in entire corpus:" + str(tfidf_matrix.sum()))
fh.write("\ni) Number and percentage of words with zero frequency in each class:" + str(zeroFreq))
fh.write("\nj) Number and percentage of word with one frequency in entire corpus:" + str(oneFreq))
fh.write("\nk) Favorite words: Christmas and tree - " + str(multiBayes.predict_log_proba(my_word_matrix)))


# Step 10 -----------------------------------------------------------------------------------------------------------
multiBayes = MultinomialNB(alpha=0.9)
multiBayes.fit(vectorizer.transform(X_train), y_train)
y_predict = multiBayes.predict(vectorizer.transform(X_test))

word_per_class = [0, 0, 0, 0, 0]
zeroFreq = [0, 0, 0, 0, 0]
oneFreq = 0
for category in range(0, len(files.target_names)):
    feature = multiBayes.feature_count_[category]
    for count in feature:
        word_per_class[category] += count
        if count == 0:
            zeroFreq[category] += 1
        if count == 1:
            oneFreq += 1

fh.write("\n\n********** MultinomialNB smoothing 0.9 **********\n")
fh.write(str(dict(zip(labels_str, counts))))
fh.write("\nb) Confusion Matrix:\n" + str(confusion_matrix(y_test, y_predict)))
fh.write("\n\nc) Classification Report: \n" + str(classification_report(y_test, y_predict, target_names=labels_str)))
fh.write("\nd) Accuracy:" + str(accuracy_score(y_test, y_predict)))
fh.write("\n   Macro F1:" + str(f1_score(y_test, y_predict, average='macro')))
fh.write("\n   Weighted F1:" + str(f1_score(y_test, y_predict, average='weighted')))
fh.write("\ne) Prior probabilities:" + str(multiBayes.class_log_prior_))  # not sure
fh.write("\nf) Vocabulary Size:" + str(len(vectorizer.vocabulary_)))
fh.write("\ng) Number of word-tokens in each class:" + str(word_per_class))  # multiBayes.class_count_
fh.write("\nh) Number of word-tokens in entire corpus:" + str(tfidf_matrix.sum()))
fh.write("\ni) Number and percentage of words with zero frequency in each class:" + str(zeroFreq))
fh.write("\nj) Number and percentage of word with one frequency in entire corpus:" + str(oneFreq))
fh.write("\nk) Favorite words: Christmas and tree - " + str(multiBayes.predict_log_proba(my_word_matrix)))

fh.close()