-
Notifications
You must be signed in to change notification settings - Fork 0
/
fyp_structured_sentiment_analysis.py
272 lines (189 loc) · 9.2 KB
/
fyp_structured_sentiment_analysis.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
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
# -*- coding: utf-8 -*-
"""FYP Structured sentiment analysis.ipynb
Automatically generated by Colaboratory.
Original file is located at
https://colab.research.google.com/drive/1QaQA8EPvuV2CpKmEPft7TxgyQ53Xpp4d
"""
!pip3 install torch==1.10.0+cu113 torchvision==0.11.1+cu113 torchaudio===0.10.0+cu113 -f https://download.pytorch.org/whl/cu113/torch_stable.html
!pip install transformers pandas numpy
!pip3 install nltk
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
import pandas as pd
import numpy as np
import nltk
"""Uploading train data.csv file """
from google.colab import files
uploaded = files.upload()
"""Pre processing test data
"""
dftest = pd.read_csv("Test_Data.csv")
dftest.info()
dftest.drop(dftest.iloc[:, 36:99], inplace = True, axis = 1)
# dftest.info()
dftest = dftest.drop(['sent_id','opinions/0/Target/1/0', 'opinions/0/Polar_expression/1/0','opinions/1/Target/1/0', 'opinions/1/Polar_expression/1/0','opinions/2/Target/1/0','opinions/2/Target/1/0','opinions/2/Polar_expression/1/0', 'opinions/3/Target/1/0','opinions/3/Polar_expression/1/0', 'opinions/4/Target/1/0','opinions/4/Polar_expression/1/0','opinions/12/Intensity','opinions/0/Source/1/0','opinions/1/Source/1/0'], axis = 1)
dftest.info()
"""Pre processing train data"""
df = pd.read_csv("Train Data.csv")
df.info()
url = 'https://raw.githubusercontent.com/anasamin26/fyp/main/Train_Data.csv'
df = pd.read_csv(url)
df.drop(df.iloc[:, 36:99], inplace = True, axis = 1)
# df.value_count()
df = df.drop(['sent_id','opinions/0/Target/1/0', 'opinions/0/Polar_expression/1/0','opinions/1/Target/1/0', 'opinions/1/Polar_expression/1/0','opinions/2/Target/1/0','opinions/2/Target/1/0','opinions/2/Polar_expression/1/0', 'opinions/3/Target/1/0','opinions/3/Polar_expression/1/0', 'opinions/4/Target/1/0','opinions/0/Source/1/0','opinions/1/Source/1/0'], axis = 1)
df.info()
# df.drop(df.iloc[:, 23:48], inplace = True, axis = 1)
# df.drop(['opinions/16/Intensity'], inplace = True, axis = 1)
df['opinions/0/Polarity']
"""Label Encoding"""
from sklearn import preprocessing
mappin = {'Positive':1, 'Negative':0, "NaN":-1}
df['opinions/0/Polarity'] = df['opinions/0/Polarity'].fillna(3)
df['opinions/0/Polarity'] = df['opinions/0/Polarity'].replace("Positive", 1)
df['opinions/0/Polarity'] = df['opinions/0/Polarity'].replace("Negative", 0)
"""# **Naive Bayes Classifier**"""
import pandas as pd
from sklearn.model_selection import train_test_split
import joblib
import pickle
from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB
vec = TfidfVectorizer(stop_words='english')
inputTexts = vec.fit_transform(df['text']).toarray()
pickle.dump(vec.vocabulary_,open("vec_vocab.pkl","wb"))
joblib.dump(vec, 'tfidf_vocab.pkl')
files.download('tfidf_vocab.pkl')
vec = CountVectorizer(stop_words='english')
inputTexts = vec.fit_transform(df['text']).toarray()
pickle.dump(vec.vocabulary_,open("vec_vocab.pkl","wb"))
# joblib.dump(vec, 'CountVectorizer.pkl')
files.download('vec_vocab.pkl')
x, x_test, y, y_test = train_test_split(inputTexts,df['opinions/0/Polarity'], test_size=0.25, random_state=42)
naiveBayes = MultinomialNB()
naiveBayes.fit(x, y)
naiveBayes.score(x_test,y_test)
print(naiveBayes.predict(vec.transform(['This was not a good idea']))[0])
import joblib
joblib.dump(naiveBayes, 'naiveBayes.pkl')
files.download('naiveBayes.pkl')
"""# **LSTM Classifier**"""
from tensorflow.keras.preprocessing.text import Tokenizer
from sklearn.preprocessing import LabelBinarizer
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import LSTM, Activation, Dropout, Dense, Input
from tensorflow.keras.layers import SpatialDropout1D
from tensorflow.keras.layers import Embedding
import tensorflow as tf
from keras.models import Model
import string
import re
import keras
from sklearn.model_selection import train_test_split
import keras.optimizers
input_df = df[['text','opinions/0/Polarity']]
input_df = df[df['opinions/0/Polarity'] != 3]
sentiment_label = df['opinions/0/Polarity'].factorize()
# df['opinions/0/Polarity']
stopwords = [ "a", "about", "above", "after", "again", "against", "all", "am", "an", "and", "any", "are", "as", "at", "be", "because",
"been", "before", "being", "below", "between", "both", "but", "by", "could", "did", "do", "does", "doing", "down", "during",
"each", "few", "for", "from", "further", "had", "has", "have", "having", "he", "he'd", "he'll", "he's", "her", "here",
"here's", "hers", "herself", "him", "himself", "his", "how", "how's", "i", "i'd", "i'll", "i'm", "i've", "if", "in", "into",
"is", "it", "it's", "its", "itself", "let's", "me", "more", "most", "my", "myself", "nor", "of", "on", "once", "only", "or",
"other", "ought", "our", "ours", "ourselves", "out", "over", "own", "same", "she", "she'd", "she'll", "she's", "should",
"so", "some", "such", "than", "that", "that's", "the", "their", "theirs", "them", "themselves", "then", "there", "there's",
"these", "they", "they'd", "they'll", "they're", "they've", "this", "those", "through", "to", "too", "under", "until", "up",
"very", "was", "we", "we'd", "we'll", "we're", "we've", "were", "what", "what's", "when", "when's", "where", "where's",
"which", "while", "who", "who's", "whom", "why", "why's", "with", "would", "you", "you'd", "you'll", "you're", "you've",
"your", "yours", "yourself", "yourselves" ]
def remove_stopwords(df):
df['review without stopwords'] = df['text'].apply(lambda x : ' '.join([word for word in x.split() if word not in (stopwords)]))
return df
def remove_tags(string):
result = re.sub('<.*?>','',string)
return result
data_without_stopwords = remove_stopwords(df)
data_without_stopwords['clean_review']= data_without_stopwords['review without stopwords'].apply(lambda cw : remove_tags(cw))
data_without_stopwords['clean_review'] = data_without_stopwords['clean_review'].str.replace('[{}]'.format(string.punctuation), ' ')
reviews_list = []
for i in range(len(df['text'])):
reviews_list.append(df['text'][i])
sentiment = data_without_stopwords['opinions/0/Polarity']
counter = 0
for item in reviews_list:
counter+=1
print(counter)
print(sentiment.size)
tokenizer = Tokenizer(num_words=5000)
tokenizer.fit_on_texts(reviews_list)
words_to_index = tokenizer.word_index
def read_glove_vector(glove_vec):
with open(glove_vec, 'r', encoding='UTF-8') as f:
words = set()
word_to_vec_map = {}
for line in f:
w_line = line.split()
curr_word = w_line[0]
word_to_vec_map[curr_word] = np.array(w_line[1:], dtype=np.float64)
return word_to_vec_map
word_to_vec_map = read_glove_vector('glove.6B.50d.txt')
maxLen = 150
vocab_len = len(words_to_index)
embed_vector_len = word_to_vec_map['moon'].shape[0]
emb_matrix = np.zeros((vocab_len, embed_vector_len))
for word, index in words_to_index.items():
embedding_vector = word_to_vec_map.get(word)
if embedding_vector is not None:
emb_matrix[index-1, :] = embedding_vector
embedding_layer = Embedding(input_dim=vocab_len, output_dim=embed_vector_len, input_length=maxLen, weights = [emb_matrix], trainable=False)
def imdb_rating(input_shape):
X_indices = Input(input_shape)
embeddings = embedding_layer(X_indices)
X = LSTM(128, return_sequences=True)(embeddings)
X = Dropout(0.6)(X)
X = LSTM(128, return_sequences=True)(X)
X = Dropout(0.6)(X)
X = LSTM(128)(X)
X = Dense(1, activation='sigmoid')(X)
model = Model(inputs=X_indices, outputs=X)
return model
counter = 0
for item in reviews_list:
counter+=1
print(counter)
X_train_indices = tokenizer.texts_to_sequences(reviews_list)
X_train_indices = pad_sequences(X_train_indices, maxlen=150, padding='post')
X_train_indices.shape
sentiment
print(X_train_indices.shape)
print(sentiment.shape)
adam = tf.keras.optimizers.Adam(learning_rate = 0.0001)
model=imdb_rating(X_train_indices.shape)
model.compile(optimizer=adam, loss='binary_crossentropy', metrics=['accuracy'])
model.fit(X_train_indices, sentiment, batch_size=64, epochs=15)
input = input_df.text.values
tokenizer = Tokenizer(num_words=5000)
tokenizer.fit_on_texts(input)
vocab_size = len(tokenizer.word_index) + 1
encoded_docs = tokenizer.texts_to_sequences(input)
padded_sequence = pad_sequences(encoded_docs, maxlen=150)
embedding_vector_length = 32
model = Sequential()
model.add(Embedding(vocab_size, embedding_vector_length,
input_length=150) )
model.add(SpatialDropout1D(0.25))
model.add(LSTM(50, dropout=0.5, recurrent_dropout=0.5))
model.add(Dropout(0.2))
model.add(Dense(1, activation='sigmoid'))
model.compile(loss='binary_crossentropy',optimizer='adam',
metrics=['accuracy'])
print(model.summary())
history = model.fit(padded_sequence,sentiment_label[0],
validation_split=0.2, epochs=20, batch_size=32)
test_word ="This is not good"
tw = tokenizer.texts_to_sequences([test_word])
tw = pad_sequences(tw,maxlen=150)
prediction = int(model.predict(tw).round().item())
sentiment_label[1][prediction]
"""# **Logistic Regression**"""
from sklearn.linear_model import LogisticRegression