/
start_nb.py
212 lines (161 loc) · 6.37 KB
/
start_nb.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
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
from sklearn.naive_bayes import MultinomialNB
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.feature_extraction.text import TfidfTransformer
from sklearn import cross_validation
from sklearn import metrics
import helpers
import numpy as np
import operator
class Start_NB(object):
""" Class for Naive Bayes classification
"""
classifier = None # Classifier of object
transformer = None # Transformer for data
vectorizer = None # Vectorizer for data
traintweets = []
train_classes = []
testtweets = []
test_classes = []
def __init__(self, pr_array, mode, tweetclass, testmode, tuplebows, ngrams, crossvalidation):
""" Initialize items """
self.pr_array = pr_array
self.tweetclass = tweetclass
self.testmode = testmode
self.CROSS_VALIDATION = crossvalidation
self.posbow, self.negbow = tuplebows
self.ngrams = ngrams
self.mode = mode
def start_classification(self, mode, new_data, allwords, fitclassifier):
""" Start classification of twitter using classifier. New_data is array of tweets divided in tokens"""
self.train_tweetclasses, self.train_vectors = self.nb_create_traintestdata(self.pr_array)
if (not fitclassifier):
self.classifier = MultinomialNB()
self.classifier.fit(self.train_vectors, self.train_tweetclasses)
if '--debug' in mode:
self.dump_classifier("classifiertest_nb.txt")
else:
self.load_classifier("classifiertest_nb.txt")
new_data_scaled = self.nb_create_inputdata(new_data, allwords)
y_pred = self.classifier.predict(new_data_scaled)
return y_pred
def start_naivebayes_evaluation(self, mode, minborder, maxborder, lenbow):
""" Start classification training of Naive Bayes"""
self.transformer = None # Reset transformer
self.vectorizer = None # Reset transformer
allwords = False
if 'allwords' in self.mode:
allwords = True
self.train_tweetclasses, self.train_vectors = self.nb_create_traintestdata(self.pr_array, allwords=allwords)
# Run Naive Bayes
results = self.run_naivebayes_evaluation(self.train_vectors, np.array(self.train_tweetclasses), self.CROSS_VALIDATION)
return results
"""
def start_classification(self, new_data):
""" """Start classification of twitter using classifier. New_data is array of tweets divided in tokens""""""
new_data_scaled = self.nb_create_vectorarray(new_data, self.scaler)
y_pred = self.classifier.predict(np.array(new_data_scaled))
return y_pred
"""
def run_naivebayes_evaluation(self, inputdata, outputdata, k):
""" Fit Naive Bayes Classification on train set with cross validation.
Run Naive Bayes Classificaiton on test set. Return results
"""
###print "** Fitting Naive Bayes classifier.."
# Cross validation
cv = cross_validation.KFold(inputdata.shape[0], n_folds=k, indices=True)
cv_naivebayes = []
f1_scores = []
for traincv, testcv in cv:
clf_cv = MultinomialNB()
clf_cv.fit(inputdata[traincv], outputdata[traincv])
y_pred_cv = clf_cv.predict(inputdata[testcv])
f1 = metrics.f1_score(outputdata[testcv], y_pred_cv, pos_label=0)
f1_scores.append(f1)
#TODO: NEEDED? self.classifier = clf_cv
print "score average: %s" + str(np.mean(f1_scores))
average_score =np.mean(f1_scores)
tuples = (average_score, f1_scores)
return (tuples, 'N.A.', 'N.A.')
def nb_create_traintestdata(self, array, **kwargs):
""" Creates dataset needed for training/testing of Naive Bayes"""
allwords = kwargs.get('allwords', False)
classes = self.tweetclass.values()
inputdata = self.nb_create_inputdata(array, allwords)
return (classes, inputdata)
def nb_create_inputdata(self, tweets, allwords):
""" Create inputdata for Naive Bayes classifier. Return data
"""
# Select BOW type
if ('posneg' in self.mode):
bow = dict(self.posbow.items() + self.negbow.items())
if ('pos1' in self.mode):
bow = self.posbow
if ('neg1' in self.mode):
bow = self.negbow
inputdata = []
if ( allwords ):
for item in tweets:
inputdata.append(' '.join(item))
else:
for item in tweets:
bowtweet = self.get_bowtweet(item, bow)
inputdata.append(bowtweet)
# Convert collection to matrix of token counts
if self.vectorizer is None:
vectorizer = CountVectorizer(ngram_range=(self.ngrams[0],self.ngrams[len(self.ngrams)-1]))
self.vectorizer = vectorizer.fit(inputdata)
X_train_counts = self.vectorizer.transform(inputdata)
# Transform count matrix to normalized tfidf representation
self.transformer = None
if self.transformer is None:
tfidf_transformer = TfidfTransformer()
self.transformer = tfidf_transformer.fit(X_train_counts)
inputdata_fitted = self.transformer.transform(X_train_counts)
return inputdata_fitted
def get_bowtweet(self, tweet, bow):
""" Get modified tweet with only words in bow"""
tuple_array = self.splitbow(bow)
listtweets = []
# Set default to False
for i in range(0, len(self.ngrams)):
listtweet = [False]*len(tweet)
listtweets.append(listtweet)
# Get array of Booleans for words occuring in BOW
for index_n, ngram in enumerate(self.ngrams):
for index_t, word in enumerate(tweet):
wordarray = []
for item in tweet[index_t:index_t+ngram]:
word = ''.join([x for x in item if ord(x) <128]) # Avoid problems with ascii values
wordarray.append(word)
wordstring = ' '.join(wordarray)
if wordstring in tuple_array[index_n]:
for i in range(index_t,index_t+ngram):
listtweets[index_n][i] = True
values = zip(*listtweets)
# Create new tweet according to booleans
new_tweet_array = []
for index, word in enumerate(tweet):
if ( any(values[index]) ):
new_tweet_array.append(word)
new_tweet = ' '.join(new_tweet_array)
return new_tweet
def splitbow(self, bow):
""" Split BOW in arrays of tuples with same length """
splitted_bowkeys = []
for ngram in self.ngrams:
ngram_array = []
for item in bow:
if len(item) == ngram:
ngram_array.append(item)
test = helpers.unfold_tuples_strings(ngram_array)
splitted_bowkeys.append(test)
return splitted_bowkeys
def load_classifier(self, filename):
""" Load classifier and scaler from file and set as class variables"""
(classifier, transformer, vectorizer) = helpers.read_from_file(filename)
self.classifier = classifier
self.scaler = scaler
def dump_classifier(self, filename):
""" Dump classifier and scaler to file """
dumptuple = (self.classifier, self.transformer, self.vectorizer)
helpers.dump_to_file(filename, dumptuple)