In [1]:

    

import pandas as pd
df = pd.read_csv('/home/dinesh/PycharmProjects/creditCardFraudDetection/creditcard.csv')


    

In [2]:

    

feature_cols = ['Time', 'V1', 'V2']


    

In [3]:

    

X = df.loc[:, feature_cols]


    

In [4]:

    

X.shape


    

    
Out[4]:





(284807, 3)

In [5]:

    

y = df.Class


    

In [6]:

    

y.shape


    

    
Out[6]:





(284807,)

In [7]:

    

from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)


    

In [8]:

    

import numpy as np
np.count_nonzero(y_train)


    

    
Out[8]:





343

In [9]:

    

np.size(y_train)


    

    
Out[9]:





190820

In [10]:

    

X_test.shape


    

    
Out[10]:





(93987, 3)

In [11]:

    

from imblearn.over_sampling import SMOTE


    

In [12]:

    

sm = SMOTE(kind='regular')


    

In [13]:

    

X_res, y_res = sm.fit_sample(X_train, y_train)


    

In [14]:

    

np.count_nonzero(y_res)


    

    
Out[14]:





190477

In [15]:

    

np.size(y_res)


    

    
Out[15]:





380954

In [16]:

    

from sklearn.ensemble import AdaBoostClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.tree import DecisionTreeClassifier


    

In [17]:

    

clf = RandomForestClassifier()
clf.fit(X_res, y_res)
scores = clf.score(X_test, y_test)


    

In [18]:

    

scores


    

    
Out[18]:





0.9730494642876143

In [19]:

    

clf = GaussianNB()
clf.fit(X_res, y_res)
scores = clf.score(X_test, y_test)


    

In [20]:

    

scores


    

    
Out[20]:





0.96701671507761711

In [21]:

    

clf = AdaBoostClassifier(DecisionTreeClassifier(max_depth=20, criterion='entropy', splitter='random'),
                        algorithm="SAMME.R")
clf.fit(X_res, y_res)
scores = clf.score(X_test, y_test)


    

In [22]:

    

scores


    

    
Out[22]:





0.97450711268579693