PROJECT : Spam SMS Data Analytics

WORK IN PROGRESS

import pandas as pd
import numpy as np
import csv

from sklearn.model_selection import StratifiedShuffleSplit

from sklearn.tree import DecisionTreeClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier, GradientBoostingClassifier
from sklearn.linear_model import SGDClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.svm import SVC,LinearSVC
from sklearn.multiclass import OneVsRestClassifier


from sklearn.metrics import classification_report

df = pd.read_csv("./SMSSpamCollection.csv", sep='\t', quoting=csv.QUOTE_NONE, names=["label", "message"])

df.head()

	label	message
0	ham	Go until jurong point, crazy.. Available only ...
1	ham	Ok lar... Joking wif u oni...
2	spam	Free entry in 2 a wkly comp to win FA Cup fina...
3	ham	U dun say so early hor... U c already then say...
4	ham	Nah I don't think he goes to usf, he lives aro...

df.label.value_counts()

ham     4827
spam     747
Name: label, dtype: int64

Loading NLTK (Natural Language Toolkit)

import nltk
nltk.download('all')

from nltk.tokenize import word_tokenize
from nltk.stem.wordnet import WordNetLemmatizer
from nltk.corpus import stopwords

Tokenization

def splitIntoTokens(text):
    text = text.lower()
    tokens = word_tokenize(text)
    return tokens

df['tokenized_message'] = df['message'].apply(splitIntoTokens)

Lemmatization (convert a word into its base form)

def getLemmas(tokens):
    lemmas = []
    lemmatizer = WordNetLemmatizer()
    for token in tokens:
        lemmas.append(lemmatizer.lemmatize(token))
    return lemmas

df['lemmatized_message'] = df['tokenized_message'].apply(getLemmas)

Removing Stop Words

stopWords = set(stopwords.words('english'))
def removeStopWords(lemmas):
    filteredSentence = []
    filteredSentence = ' '.join([word for word in lemmas if word not in stopWords])
    return filteredSentence

df['filtered_message'] = df['lemmatized_message'].apply(removeStopWords)

TFIDF Matrix

• The Term Document Matrix (TDM) is a matrix that contains the frequency of occurrence of terms in a collection of documents.

• In a Term Frequency Inverse Document Frequency (TFIDF) matrix, the term importance is expressed by Inverse Document Frequency (IDF)

• IDF diminishes the weight of the most commonly occurring words and increases the weightage of rare words.

from sklearn.feature_extraction.text import TfidfVectorizer

tfidfVectorizer = TfidfVectorizer(
    ngram_range=(1, 2),
    min_df = 1/df.shape[0],
    max_df = 0.7
)

tfidfModel = tfidfVectorizer.fit(df['filtered_message'])

X = tfidfModel.transform(df['filtered_message']).toarray()
y = df['label'].values

print(X.shape, y.shape)

(5574, 40373) (5574,)

Stratified Shuffle Sampling

data_gen = StratifiedShuffleSplit(n_splits=1, test_size=0.3, random_state=7)

Models Pool

classifiers = [
    DecisionTreeClassifier(),
    GaussianNB(),
    SGDClassifier(loss='modified_huber', shuffle=True),
    SVC(kernel="linear", C=0.025),
    KNeighborsClassifier(),
    OneVsRestClassifier(LinearSVC()),
    RandomForestClassifier(max_depth=5, n_estimators=100, max_features=10),
    AdaBoostClassifier(),
]

models_params = []
for model in classifiers:
    for train_index, test_index in data_gen.split(X,y):
        X_train, X_test = X[train_index], X[test_index]
        y_train, y_test = y[train_index], y[test_index]
        model.fit(X_train, y_train)

        y_pred = model.predict(X_test)
        report = classification_report(y_test, y_pred, output_dict=True)

        model_report = report['spam']
        model_report.update({
            'model' : type(model).__name__,
            'accuracy' : report['accuracy']
        })

        models_params.append(model_report)

report_df = pd.DataFrame(models_params)

report_df

	precision	recall	f1-score	support	model	accuracy
0	0.880734	0.857143	0.868778	224	DecisionTreeClassifier	0.965332
1	0.658863	0.879464	0.753346	224	GaussianNB	0.922893
2	1.000000	0.816964	0.899263	224	SGDClassifier	0.975493
3	0.000000	0.000000	0.000000	224	SVC	0.866109
4	1.000000	0.160714	0.276923	224	KNeighborsClassifier	0.887627
5	0.989130	0.812500	0.892157	224	OneVsRestClassifier	0.973700
6	0.000000	0.000000	0.000000	224	RandomForestClassifier	0.866109
7	0.943590	0.821429	0.878282	224	AdaBoostClassifier	0.969516

del report_df['support']

report_df.plot.bar(x='model')

<matplotlib.axes._subplots.AxesSubplot at 0x7f9f850c3b10>