-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathLDA.py
176 lines (139 loc) · 7.81 KB
/
LDA.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
import os
import numpy as np
import pandas as pd
import sklearn
from sklearn.decomposition import LatentDirichletAllocation
from sklearn.feature_extraction.text import TfidfVectorizer
from src.utils import most_freq_topic_words
import warnings
warnings.filterwarnings("ignore")
preprocessed_dir = os.path.join(os.curdir,'data','preprocessed')
tables_dir = os.path.join(os.curdir,'tables')
feb_tweets = pd.read_pickle(os.path.join(preprocessed_dir,'feb_tweets.pkl'))
mar_tweets = pd.read_pickle(os.path.join(preprocessed_dir,'mar_tweets.pkl'))
apr_tweets = pd.read_pickle(os.path.join(preprocessed_dir,'apr_tweets.pkl'))
# Data splitting for LDA purposes.
feb_pos = feb_tweets[feb_tweets.label == 'positive']
feb_neg = feb_tweets[feb_tweets.label == 'nigative']
mar_pos = mar_tweets[mar_tweets.label == 'positive']
mar_neg = mar_tweets[mar_tweets.label == 'nigative']
apr_pos = apr_tweets[apr_tweets.label == 'positive']
apr_neg = apr_tweets[apr_tweets.label == 'nigative']
# TF-IDF vectorizer instanciation
vectorizer_fp = TfidfVectorizer(analyzer='word',
min_df = 10,
max_df = 0.9,
stop_words='english',
token_pattern='[a-zA-Z0-9]{3,}',
)
vectorizer_fn = TfidfVectorizer(analyzer='word',
min_df = 10,
max_df = 0.9,
stop_words='english',
token_pattern='[a-zA-Z0-9]{3,}',
)
vectorizer_mp = TfidfVectorizer(analyzer='word',
min_df = 10,
max_df = 0.9,
stop_words='english',
token_pattern='[a-zA-Z0-9]{3,}',
)
vectorizer_mn = TfidfVectorizer(analyzer='word',
min_df = 10,
max_df = 0.9,
stop_words='english',
token_pattern='[a-zA-Z0-9]{3,}',
)
vectorizer_ap = TfidfVectorizer(analyzer='word',
min_df = 10,
max_df = 0.9,
stop_words='english',
token_pattern='[a-zA-Z0-9]{3,}',
)
vectorizer_an = TfidfVectorizer(analyzer='word',
min_df = 10,
max_df = 0.9,
stop_words='english',
token_pattern='[a-zA-Z0-9]{3,}',
)
# TF-IDF vectrorizer fitting
feb_pos_vec = vectorizer_fp.fit_transform(list(feb_pos['clean_stemmed']))
feb_neg_vec = vectorizer_fn.fit_transform(list(feb_neg['clean_stemmed']))
mar_pos_vec = vectorizer_mp.fit_transform(list(mar_pos['clean_stemmed']))
mar_neg_vec = vectorizer_mn.fit_transform(list(mar_neg['clean_stemmed']))
apr_pos_vec = vectorizer_ap.fit_transform(list(apr_pos['clean_stemmed']))
apr_neg_vec = vectorizer_an.fit_transform(list(apr_neg['clean_stemmed']))
# LDA model instanciation
lda_model_fp = LatentDirichletAllocation(n_components=4,
max_iter=20,
learning_method='online',
random_state=100,
batch_size=128,
evaluate_every = -1,
n_jobs = -1,
)
lda_model_fn = LatentDirichletAllocation(n_components=4,
max_iter=20,
learning_method='online',
random_state=100,
batch_size=128,
evaluate_every = -1,
n_jobs = -1,
)
lda_model_mp = LatentDirichletAllocation(n_components=4,
max_iter=20,
learning_method='online',
random_state=100,
batch_size=128,
evaluate_every = -1,
n_jobs = -1,
)
lda_model_mn = LatentDirichletAllocation(n_components=4,
max_iter=20,
learning_method='online',
random_state=100,
batch_size=128,
evaluate_every = -1,
n_jobs = -1,
)
lda_model_ap = LatentDirichletAllocation(n_components=4,
max_iter=20,
learning_method='online',
random_state=100,
batch_size=128,
evaluate_every = -1,
n_jobs = -1,
)
lda_model_an = LatentDirichletAllocation(n_components=4,
max_iter=20,
learning_method='online',
random_state=100,
batch_size=128,
evaluate_every = -1,
n_jobs = -1,
)
# LDA model fitting
feb_pos_tm = lda_model_fp.fit_transform(feb_pos_vec)
print('February positive topic modelling completed')
feb_neg_tm = lda_model_fn.fit_transform(feb_neg_vec)
print('February negative topic modelling completed')
mar_pos_tm = lda_model_mp.fit_transform(mar_pos_vec)
print('March positive topic modelling completed')
mar_neg_tm = lda_model_mn.fit_transform(mar_neg_vec)
print('march negative topic modelling completed')
apr_pos_tm = lda_model_ap.fit_transform(apr_pos_vec)
print('April positive topic modelling completed')
apr_neg_tm = lda_model_an.fit_transform(apr_neg_vec)
print('April negative topic modelling completed')
fp = most_freq_topic_words(lda_model_fp,vectorizer_fp,10)
fp.to_csv(os.path.join(tables_dir,'most-frequent-positive-words-February'))
fn = most_freq_topic_words(lda_model_fn,vectorizer_fn,10)
fn.to_csv(os.path.join(tables_dir,'most-frequent-negative-words-February'))
mp = most_freq_topic_words(lda_model_mp,vectorizer_mp,10)
mp.to_csv(os.path.join(tables_dir,'most-frequent-positive-words-March'))
mn = most_freq_topic_words(lda_model_mn,vectorizer_mn,10)
mn.to_csv(os.path.join(tables_dir,'most-frequent-negative-words-March'))
ap = most_freq_topic_words(lda_model_ap,vectorizer_ap,10)
ap.to_csv(os.path.join(tables_dir,'most-frequent-positive-words-April'))
an = most_freq_topic_words(lda_model_an,vectorizer_an,10)
an.to_csv(os.path.join(tables_dir,'most-frequent-negative-words-April'))