Jovian
⭐️
Sign In
NotebookFilesCollaborators
Version 1

Building machine learning model for spam classifier for text message

load all the required libraries

In [99]:
import numpy as np
import pandas as pd
import re
import spacy as sp
import nltk as nl
import seaborn as sns
import matplotlib.pyplot as plt

from sklearn.feature_extraction.text import CountVectorizer,TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix,accuracy_score,classification_report

load english model for spacy

In [2]:
nlp = sp.load("en_core_web_sm")

read the raw data

In [5]:
df = pd.read_table("SMSSpamCollection.txt",names=["labels","message"])
df.head(2)
/Users/vigneshkarthick/anaconda3/lib/python3.7/site-packages/ipykernel_launcher.py:1: FutureWarning: read_table is deprecated, use read_csv instead, passing sep='\t'. """Entry point for launching an IPython kernel.
Out[5]:
In [6]:
df = pd.read_csv("SMSSpamCollection.txt",sep="\t",names=["labels","message"])
df.head(2)
Out[6]:

Analyze the data

In [8]:
df.shape
Out[8]:
(5572, 2)
In [11]:
df.message[:5]
Out[11]:
0    Go until jurong point, crazy.. Available only ...
1                        Ok lar... Joking wif u oni...
2    Free entry in 2 a wkly comp to win FA Cup fina...
3    U dun say so early hor... U c already then say...
4    Nah I don't think he goes to usf, he lives aro...
Name: message, dtype: object

Perform preprocessing on raw data and apply function over message

  • make lower
  • remove special chars
  • create nlp object
  • remove stop words and punctuation using list comprehension
  • remove spaces again
  • convert to sentence again by joining the words
  • Remove bigger spaces which got created after joining
  • return the clean sentence
In [26]:
def textPreprocessor(msg):
    msg = msg.lower()
    msg = re.sub("[^\w]"," ",msg)
    obj = nlp(msg)
    msg_wo_sw_pun = [wd.text for wd in obj if wd.is_stop == False and wd.is_punct == False]
    sent = " ".join(msg_wo_sw_pun)
    clean_sent = re.sub("\s+"," ",sent)
    return clean_sent
# textPreprocessor(df.message[200])
In [27]:
df.message = df.message.apply(textPreprocessor)
In [30]:
df.message[:10]
Out[30]:
0    jurong point crazy available bugis n great wor...
1                             ok lar joking wif u oni 
2    free entry 2 wkly comp win fa cup final tkts 2...
3                                 u dun early hor u c 
4                       nah don t think goes usf lives
5    freemsg hey darling s 3 week s word d like fun...
6            brother like speak treat like aids patent
7    request melle melle oru minnaminunginte nurung...
8    winner valued network customer selected receiv...
9    mobile 11 months u r entitled update latest co...
Name: message, dtype: object

Label 0/1 on label column

In [32]:
df.labels[:5]
Out[32]:
0     ham
1     ham
2    spam
3     ham
4     ham
Name: labels, dtype: object
In [34]:
df.labels.replace({"ham":0,"spam":1},inplace=True)
In [35]:
df.labels[:5]
Out[35]:
0    0
1    0
2    1
3    0
4    0
Name: labels, dtype: int64

Create train and test data

In [ ]:
# train = df.message[:-200]
# test = df.message[-200:]
In [36]:
X = df.message# X is my independent data will be used for training and testing
y = df.labels# y is my dependent(target) data will be used for training and testing
In [37]:
X.shape
Out[37]:
(5572,)
In [38]:
y.shape
Out[38]:
(5572,)
In [40]:
X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=.2,random_state=6) # random_state is used for reproducibility
In [51]:
X_train.shape
Out[51]:
(4457,)
In [52]:
X_test.shape
Out[52]:
(1115,)
In [54]:
y_train.shape
Out[54]:
(4457,)
In [53]:
y_test.shape
Out[53]:
(1115,)
In [50]:
# X_train,X_test,y_train,y_test = train_test_split(X,y,test_size=.2,random_state=7)
# X_train

Create BOW (bag of words) using Count Vectorizer

In [71]:
cv = CountVectorizer()
cv
Out[71]:
CountVectorizer(analyzer='word', binary=False, decode_error='strict',
        dtype=<class 'numpy.int64'>, encoding='utf-8', input='content',
        lowercase=True, max_df=1.0, max_features=None, min_df=1,
        ngram_range=(1, 1), preprocessor=None, stop_words=None,
        strip_accents=None, token_pattern='(?u)\\b\\w\\w+\\b',
        tokenizer=None, vocabulary=None)
In [83]:
cv_train_dtm = cv.fit_transform(X_train)
cv_train_dtm
Out[83]:
<4457x7488 sparse matrix of type '<class 'numpy.int64'>'
	with 34567 stored elements in Compressed Sparse Row format>
In [84]:
cv_train_dtm.toarray() # to see sparse matrix
Out[84]:
array([[0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0],
       ...,
       [0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0],
       [0, 0, 0, ..., 0, 0, 0]], dtype=int64)
In [85]:
print(cv_train_dtm.tocoo()) # to see co-ordinates of non-zero values
(0, 5716) 1 (0, 3221) 1 (0, 1003) 1 (0, 5994) 1 (0, 6441) 1 (0, 4491) 1 (0, 2188) 1 (0, 753) 1 (0, 2254) 1 (1, 2848) 1 (1, 1460) 1 (1, 3859) 1 (1, 4704) 1 (2, 2390) 1 (2, 6472) 1 (2, 4774) 1 (3, 4641) 1 (3, 3601) 1 (3, 7152) 1 (4, 6547) 1 (4, 6708) 1 (4, 3744) 1 (4, 853) 1 (4, 6536) 1 (4, 2611) 1 : : (4454, 4144) 1 (4454, 4544) 1 (4454, 3095) 1 (4455, 549) 1 (4455, 5023) 1 (4455, 4940) 1 (4455, 5788) 1 (4455, 5726) 1 (4455, 3347) 1 (4455, 2452) 1 (4455, 4218) 1 (4455, 4000) 1 (4455, 979) 1 (4455, 6834) 1 (4456, 4434) 1 (4456, 1926) 1 (4456, 4329) 1 (4456, 3953) 1 (4456, 4786) 1 (4456, 3149) 1 (4456, 1975) 1 (4456, 3405) 1 (4456, 6730) 1 (4456, 2923) 1 (4456, 3859) 1
In [82]:
cv_test_dtm = cv.transform(X_test)
In [62]:
# cv2 = CountVectorizer(ngram_range=(1,2))
# cv2.fit_transform(X)
In [67]:
# cv2.get_feature_names()

Create MultinomialNB() instance for training and testing

https://monkeylearn.com/blog/practical-explanation-naive-bayes-classifier/

In [86]:
nb = MultinomialNB()
In [88]:
nb.fit(cv_train_dtm,y_train)
Out[88]:
MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True)

Predict the labels on test data

In [90]:
y_pred = nb.predict(cv_test_dtm)
y_pred
Out[90]:
array([0, 0, 0, ..., 0, 0, 0])

Confusion matrix

In [91]:
confusion_matrix(y_test,y_pred)
Out[91]:
array([[963,   5],
       [ 10, 137]])

Visualize confusion matrix using heatmap

In [104]:
sns.heatmap(confusion_matrix(y_test,y_pred),annot=True,fmt="d")
Out[104]:
<matplotlib.axes._subplots.AxesSubplot at 0x1a7f052908>
Notebook Image

Accuracy score

In [97]:
accuracy_score(y_test,y_pred)*100
Out[97]:
98.65470852017937

Create BOW (bag of words) using tfidf-Vectorizer

In [105]:
tfidf = TfidfVectorizer()
In [118]:
tf_train = tfidf.fit_transform(X_train[:5])
tf_train
Out[118]:
<5x41 sparse matrix of type '<class 'numpy.float64'>'
	with 41 stored elements in Compressed Sparse Row format>
In [119]:
tf_train.toarray()
Out[119]:
array([[0.33333333, 0.        , 0.        , 0.33333333, 0.        ,
        0.        , 0.33333333, 0.33333333, 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.33333333, 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.33333333, 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.33333333, 0.33333333, 0.        , 0.33333333,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        ],
       [0.        , 0.        , 0.        , 0.        , 0.5       ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.5       , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.5       ,
        0.        , 0.        , 0.        , 0.        , 0.5       ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        ],
       [0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.57735027, 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.57735027, 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.57735027, 0.        , 0.        , 0.        , 0.        ,
        0.        ],
       [0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.57735027, 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.57735027, 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.        , 0.        , 0.        , 0.        , 0.        ,
        0.57735027],
       [0.        , 0.21320072, 0.21320072, 0.        , 0.        ,
        0.21320072, 0.        , 0.        , 0.        , 0.21320072,
        0.21320072, 0.        , 0.21320072, 0.        , 0.21320072,
        0.21320072, 0.        , 0.21320072, 0.21320072, 0.        ,
        0.21320072, 0.21320072, 0.        , 0.        , 0.        ,
        0.        , 0.21320072, 0.21320072, 0.21320072, 0.21320072,
        0.21320072, 0.        , 0.        , 0.21320072, 0.        ,
        0.        , 0.21320072, 0.21320072, 0.21320072, 0.21320072,
        0.        ]])
In [114]:
tf_train.toarray()[0]
Out[114]:
0.0

Serializing/De-serializing the model

In [122]:
from joblib import dump,load
In [123]:
dump(nb,"nb.joblib")
Out[123]:
['nb.joblib']
In [124]:
mod = load("nb.joblib")
In [127]:
y_pred = mod.predict(cv_test_dtm)
y_pred
Out[127]:
array([0, 0, 0, ..., 0, 0, 0])
In [128]:
accuracy_score(y_test,y_pred)*100
Out[128]:
98.65470852017937