-
Notifications
You must be signed in to change notification settings - Fork 0
/
Spam Detection .py
313 lines (148 loc) · 4.94 KB
/
Spam Detection .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
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
#!/usr/bin/env python
# coding: utf-8
# In[67]:
# SMS Spam Classification
# Create a model that can predict whether the given sms is a spam or ham sms.
# In[68]:
import numpy as np
import pandas as pd
# In[69]:
data = pd.read_csv('C:\Datasets\SMSSpamCollection', sep='\t', names=['label','message'])
# In[70]:
data.info()
# In[71]:
data.head()
# In[72]:
# Balanced and Unbalanced Dataset in Classification
# Balanced Dataset -> no of observations for each unique label is same.
# e.g. n(ham) = n(spam) ---> Balanced Dataset
# Unbalanced Dataset -> no of observations of each unique label is different.
# e.g. n(ham) not equal to n(spam) ---> Unbalanced Dataset
data.label.value_counts()
# In[73]:
# The given dataset is an Unbalanced Dataset.
# In[74]:
# Preprocess Feature Column since given feature column (message) is Pure String
# 1. Perform Text Preprocessing
# 2. Create BoW
# 3. Apply TF-IDF on BoW
# In[75]:
# 1. Perform Text Preprocessing
#
# a. Remove Punctuations
# b. Convert Sentences to Words
# c. Remove Stopwords
# d. Normalize the words
# In[76]:
import nltk
from nltk.corpus import stopwords
import string
def textPreprocessor(feature):
# a. Remove Punctuations
removePunctuations = [character for character in feature if character not in string.punctuation]
sentencesWithoutPunct = ''.join(removePunctuations)
# b. Convert Sentences to Words -- Tokenization
words = sentencesWithoutPunct.split(" ")
# c. Normalize the words
wordNormalized = [ word.lower() for word in words ]
# d. Remove Stopwords
finalWords = [word for word in wordNormalized if word not in stopwords.words('english')]
return finalWords
# In[77]:
# Seperate data as features and label
features = data.iloc[:,[1]].values
features
# In[78]:
label = data.iloc[:,[0]].values
label
# In[79]:
# Creating BOW -- scikit-learn
# feature ---> textPreprocessor ---> Creating BOW( Vocab , Contigency Matrix)
from sklearn.feature_extraction.text import CountVectorizer
wordVector = CountVectorizer(analyzer=textPreprocessor)
# Building the Vocabulary
finalWordVectorVocab = wordVector.fit(features)
# In[80]:
finalWordVectorVocab.vocabulary_
# In[81]:
# Building BOW
bagOfWords = finalWordVectorVocab.transform(features)
# In[82]:
demo = bagOfWords.toarray()
demo
# In[83]:
len(finalWordVectorVocab.vocabulary_)
# In[84]:
demo.shape
# In[85]:
# Applying TFIDF on BOW
from sklearn.feature_extraction.text import TfidfTransformer
tfIdfObject = TfidfTransformer().fit(bagOfWords)
# In[86]:
finalFeatureArray = tfIdfObject.transform(bagOfWords)
# In[87]:
finalFeatureArray
# In[88]:
# import sys
# import numpy as np
# np.set_printoptions(threshold=sys.maxsize)
# demo1 = finalFeatureArray.toarray()
# demo1
# In[89]:
# pd.DataFrame(finalFeatureArray.toarray()).to_csv('TFIDF.csv', index=False)
# In[90]:
# Train Test Split
from sklearn.model_selection import train_test_split
X_train,X_test,y_train,y_test = train_test_split(finalFeatureArray,
label,
test_size=0.2,
random_state=6)
# In[91]:
# Building the model
from sklearn.linear_model import LogisticRegression
model = LogisticRegression()
model.fit(X_train,y_train)
# In[92]:
# Model score
model.score(X_train,y_train)
# In[93]:
model.score(X_test,y_test)
# In[94]:
# Creating a function for the same
import warnings
warnings.filterwarnings("ignore")
for i in range(1,101):
X_train,X_test,y_train,y_test = train_test_split(finalFeatureArray,
label,
test_size=0.2,
random_state=i)
model = LogisticRegression()
model.fit(X_train,y_train)
trainScore = model.score(X_train,y_train)
testScore = model.score(X_test,y_test)
if testScore > trainScore and testScore > 0.9:
print("Testing {} , Training {}, RS {}".format(testScore,trainScore,i))
# In[95]:
# The dataset is an unbalanced dataset, we need to check for precision and recall
from sklearn.metrics import confusion_matrix
confusion_matrix(label , model.predict(finalFeatureArray))
# In[96]:
from sklearn.metrics import classification_report
print(classification_report(label , model.predict(finalFeatureArray)))
# In[97]:
# Avg of Precision(Spam ) and Recall(Ham) --> 1 which is greater than CL --- Thus ACCEPTABLE !!!!
# In[98]:
#Deployment Check
smsInput = input("Enter SMS: ")
#Preprocess
preProcessedFeature = textPreprocessor(smsInput)
#BOW
bowFeature = finalWordVectorVocab.transform(preProcessedFeature)
#TFIDF
tfIDFFeature = tfIdfObject.transform(bowFeature)
#Pred
predLabel = model.predict(tfIDFFeature)[0]
print("Given SMS is {}".format(predLabel))
# In[99]:
# Thank You :)
# In[ ]: