-
Notifications
You must be signed in to change notification settings - Fork 0
/
decisiontree.py
172 lines (156 loc) · 6.13 KB
/
decisiontree.py
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
import utils
import random
import numpy as np
from sklearn.tree import DecisionTreeClassifier
from scipy.sparse import lil_matrix
from sklearn.feature_extraction.text import TfidfTransformer
import time
# Performs classification using Decision Tree.
FREQ_DIST_FILE = '../train-processed-freqdist.pkl'
BI_FREQ_DIST_FILE = '../train-processed-freqdist-bi.pkl'
TRAIN_PROCESSED_FILE = '../train-processed.csv'
TEST_PROCESSED_FILE = '../test-processed.csv'
# True while training.
TRAIN = True
UNIGRAM_SIZE = 15000
VOCAB_SIZE = UNIGRAM_SIZE
# If using bigrams.
USE_BIGRAMS = False
if USE_BIGRAMS:
BIGRAM_SIZE = 10000
VOCAB_SIZE = UNIGRAM_SIZE + BIGRAM_SIZE
FEAT_TYPE = 'frequency'
def get_feature_vector(tweet):
uni_feature_vector = []
bi_feature_vector = []
words = tweet.split()
for i in xrange(len(words) - 1):
word = words[i]
next_word = words[i + 1]
if unigrams.get(word):
uni_feature_vector.append(word)
if USE_BIGRAMS:
if bigrams.get((word, next_word)):
bi_feature_vector.append((word, next_word))
if len(words) >= 1:
if unigrams.get(words[-1]):
uni_feature_vector.append(words[-1])
return uni_feature_vector, bi_feature_vector
def extract_features(tweets, batch_size=500, test_file=True, feat_type='presence'):
num_batches = int(np.ceil(len(tweets) / float(batch_size)))
for i in xrange(num_batches):
batch = tweets[i * batch_size: (i + 1) * batch_size]
features = lil_matrix((batch_size, VOCAB_SIZE))
labels = np.zeros(batch_size)
for j, tweet in enumerate(batch):
if test_file:
tweet_words = tweet[1][0]
tweet_bigrams = tweet[1][1]
else:
tweet_words = tweet[2][0]
tweet_bigrams = tweet[2][1]
labels[j] = tweet[1]
if feat_type == 'presence':
tweet_words = set(tweet_words)
tweet_bigrams = set(tweet_bigrams)
for word in tweet_words:
idx = unigrams.get(word)
if idx:
features[j, idx] += 1
if USE_BIGRAMS:
for bigram in tweet_bigrams:
idx = bigrams.get(bigram)
if idx:
features[j, UNIGRAM_SIZE + idx] += 1
yield features, labels
def apply_tf_idf(X):
"""
Fits X for TF-IDF vectorization and returns the transformer.
"""
transformer = TfidfTransformer(smooth_idf=True, sublinear_tf=True, use_idf=True)
transformer.fit(X)
return transformer
def process_tweets(csv_file, test_file=True):
"""Returns a list of tuples of type (tweet_id, feature_vector)
or (tweet_id, sentiment, feature_vector)
Args:
csv_file (str): Name of processed csv file generated by preprocess.py
test_file (bool, optional): If processing test file
Returns:
list: Of tuples
"""
tweets = []
print 'Generating feature vectors'
with open(csv_file, 'r') as csv:
lines = csv.readlines()
total = len(lines)
for i, line in enumerate(lines):
if test_file:
tweet_id, tweet = line.split(',')
else:
tweet_id, sentiment, tweet = line.split(',')
feature_vector = get_feature_vector(tweet)
if test_file:
tweets.append((tweet_id, feature_vector))
else:
tweets.append((tweet_id, int(sentiment), feature_vector))
utils.write_status(i + 1, total)
print '\n'
return tweets
if __name__ == '__main__':
np.random.seed(1337)
unigrams = utils.top_n_words(FREQ_DIST_FILE, UNIGRAM_SIZE)
if USE_BIGRAMS:
bigrams = utils.top_n_bigrams(BI_FREQ_DIST_FILE, BIGRAM_SIZE)
tweets = process_tweets(TRAIN_PROCESSED_FILE, test_file=False)
if TRAIN:
train_tweets, val_tweets = utils.split_data(tweets)
else:
random.shuffle(tweets)
train_tweets = tweets
del tweets
print 'Extracting features & training batches'
clf = DecisionTreeClassifier(max_depth=25)
batch_size = len(train_tweets)
i = 1
n_train_batches = int(np.ceil(len(train_tweets) / float(batch_size)))
for training_set_X, training_set_y in extract_features(train_tweets, test_file=False, feat_type=FEAT_TYPE, batch_size=batch_size):
utils.write_status(i, n_train_batches)
i += 1
if FEAT_TYPE == 'frequency':
tfidf = apply_tf_idf(training_set_X)
training_set_X = tfidf.transform(training_set_X)
clf.fit(training_set_X, training_set_y)
print '\n'
print 'Testing'
if TRAIN:
correct, total = 0, len(val_tweets)
i = 1
batch_size = len(val_tweets)
n_val_batches = int(np.ceil(len(val_tweets) / float(batch_size)))
for val_set_X, val_set_y in extract_features(val_tweets, test_file=False, feat_type=FEAT_TYPE, batch_size=batch_size):
if FEAT_TYPE == 'frequency':
val_set_X = tfidf.transform(val_set_X)
prediction = clf.predict(val_set_X)
correct += np.sum(prediction == val_set_y)
utils.write_status(i, n_val_batches)
i += 1
print '\nCorrect: %d/%d = %.4f %%' % (correct, total, correct * 100. / total)
else:
del train_tweets
test_tweets = process_tweets(TEST_PROCESSED_FILE, test_file=True)
n_test_batches = int(np.ceil(len(test_tweets) / float(batch_size)))
predictions = np.array([])
print 'Predicting batches'
i = 1
for test_set_X, _ in extract_features(test_tweets, test_file=True, feat_type=FEAT_TYPE):
if FEAT_TYPE == 'frequency':
test_set_X = tfidf.transform(test_set_X)
prediction = clf.predict(test_set_X)
predictions = np.concatenate((predictions, prediction))
utils.write_status(i, n_test_batches)
i += 1
predictions = [(str(j), int(predictions[j]))
for j in range(len(test_tweets))]
utils.save_results_to_csv(predictions, 'decisiontree.csv')
print '\nSaved to decisiontree.csv'