Welcome to Scribd. Sign in or start your free trial to enjoy unlimited e-books, audiobooks & documents.Find out more
Download
Standard view
Full view
of .
Look up keyword
Like this
2Activity
0 of .
Results for:
No results containing your search query
P. 1
A Novel Approach to Credit Card Fraud Detection Model

A Novel Approach to Credit Card Fraud Detection Model

Ratings: (0)|Views: 697|Likes:
Journal of Computing, ISSN 2151-9617, http://www.journalofcomputing.org
Journal of Computing, ISSN 2151-9617, http://www.journalofcomputing.org

More info:

Published by: Journal of Computing on Jan 22, 2011
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

01/22/2011

pdf

text

original

 
JOURNAL OF COMPUTING, VOLUME 2, ISSUE12, DECEMBER 2010, ISSN 2151-9617HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/ WWW.JOURNALOFCOMPUTING.ORG 94
A Novel Approach to Credit Card FraudDetection Model
V.Dheepa,
 
Dr.R.Dhanapal, Remigious D
Research and Development Centre, Bharathiar University, Coimbatore, Tamilnadu, India.
 
Abstract
 — Along with the great increase in credit card transactions, credit card fraud has become increasingly rampant inrecent years. In Modern day the fraud is one of the major causes of great financial losses, not only for merchants, individualclients are also affected. In this paper, clustering and outlier detection techniques are used to find the fraudulent activities. Inthe first phase, clustering is used to partition the data. In the second phase two different outlier detection algorithms are used inthe partitions separately for finding the outliers. Finally, the outliers are combined and fraudulent cases are found.
Index Terms
 — Clustering, Credit Card Fraud Detection, Outlier Detection.
——————————
 
 
——————————
1 I
NTRODUCTION
The use of credit cards is prevalent in modern day socie-ty. Credit card becomes the most popular mode of pay-ment. Detecting credit card fraud is a difficult task whenusing normal procedures, so the development of the cre-dit card fraud detection model has become of signific-ance, whether in the academic or business communityrecently.Detecting the fraud means identifying suspicious frau-dulent cases. In this paper, clustering and outlier mining isused to detect the fraudulent cases. Clustering is used togroup the similar data objects into clusters. Outliers aredefined as data that appears to be inconsistent with the restof the data. Mining of outliers is an important researchfield in the application of fraud detection. Detection of out-liers has recently gained a lot of application in many do-mains. In this model, outlier detection algorithm is em-ployed to find fraudulent transactions.
2
 
R
ELATED
W
ORK
From the work of view for preventing credit card fraud,more research works were carried out with special em-phasis on data mining. Kim and Kim have identifiedskewed distribution of data and mix of Legitimate andfraudulent transactions as the two main reasons for thecomplexity of credit card fraud detection [1]. Sam andKarl suggest a credit card fraud detection model usingBayesian networks and neural network techniques tolearn models of fraudulent credit card transactions[2].L.MuKhanov finds a credit card fraud detection modelusing Bayesian Belief Networks [3].Some Clustering-based outlier detection techniques are al-so proposed to find the Fraud detection [4].Grid based Outlier detection was also projected [5]. There is alot of research in outlier detection. Many outlier detection al-gorithms such as base on statistics [6] and distance [7, 8] aregain good application.In this paper, the method of detection outliers is used forset up a detection model, which could mine fraud transactionsas outliers.
3 F
RAUD
D
ETECTION
M
ODEL
 
Fig.1.Outline of the Fraud Detection Model
Lecturer, Department of M.Sc-InformationTechnology,Velammal Engineering College, Chennai.
Professor,
 
Department of Computer Applications,Easwari Engineering College, Chennai.
 
 
SAP Consultant,Wipro Technolgogies,Client: Apple Computers, Cupertino-California, U.S.
 
JOURNAL OF COMPUTING, VOLUME 2, ISSUE 12, DECEMBER 2010, ISSN 2151-9617HTTPS://SITES.GOOGLE.COM/SITE/JOURNALOFCOMPUTING/ WWW.JOURNALOFCOMPUTING.ORG 95
An Outline of the Fraud detection model is shown in Fig.1.First step in this method is Data Preprocessing. According toevery attribute of the transaction sample, the model doessome data preprocessing to convert all of them to numericalattribute. The second and Third step shows the detection me-thods. The Final step concludes the fraudulent cases.
4
 
D
ETECTION
M
ETHODS
 
4.1 Clustering
Clustering helps in grouping the data into similar clusters thathelps in uncomplicated retrieval of data [9].Cluster analysis isa technique for breaking data down into related componentsin such a way that patterns and order becomes visible[10].Clustering techniques are known as “unsupervised learn-ing” because there is no class to be predicted. The main goal ofclustering data is to find common patterns or to group similarcases in the data. In this paper, an efficient cluster based parti-tioning algorithm is used. This divides the data in specifiednumber of partitions which shows the partitions of dense re-gions and sparse regions. The K-means algorithm is applied tocluster the data, which find out the sparse region and denseregion. K-means clustering is a method of cluster analysiswhich aims to partition n observations into k clusters in whicheach observation belongs to the cluster with the nearest mean.The cluster mean of Ki={t
i1
,t
i2
,….,t
im
} is defined asThe partitioned data region is shown in Fig. 2.Fig.2.Partitioned Data Region using clustering
4.2 Outlier Detection
Outliers are defined as the data which deviate from other ob-servations so as to arouse suspicion that they were generatedby a different mechanism. In this study, two outlier detectionmethods are employed for two different regions.For dense regions, the generalized extreme studentizeddeviate (GESD) is used.This method identifies outlierswhich are data samples that vary significantly from themajority of the data. The degree to which each outlierdeviates from the remainder of the data indicates theseverity of the abnormal activity denoted by that outlier.It is based on the mean and standard deviation of the dataobserved. In this method, two parameters are used, an up-per bound, N
u
, on the number of potential outliers and theprobability,
α
, of incorrectly declaring one or more outlierswhen no outliers exist. N
u
<= 1/ 2 (N - 1), where N is thenumber of samples. This method works well for the densi-ty region.For sparse regions, the Q-test outlier detection method isused. First, the set of data are arranged in ascending order.Then the experimental Q-value is calculated. It is the ratiowhich shows the variation of the suspect value from itsnearest one divided by the range of the values. The Q valueis compared with the critical Q- value (Q_crit).Critical Q-value is defined with confidence level. If Q>Q
crit
then thesuspect value can be consider as outlier. This method per-forms well with small sample sizes.All the detected outliers from the regions are com-bined. They are ranked based on the severity of the outliersand concluded as fraudulent cases.
5
 
D
ISCUSSIONS
In this detection Model, the unsupervised approach is used.The unknown frauds are easily found by using this approach.The models based on supervised approach must have thelabeled data for both normal data and anomolies.It is onlyable to detect frauds of a type which has previously occurred.In contrast, unsupervised methods don’t make use of labeledrecords. It detects the changes in behavior or unusual Transac-tions. Unsupervised learning is a feasible method to learn thelarge and more complex model.
 
In this process, the Clustering and Outlier Detectionmethods are worn to find the Outliers. By applying thealgorithms in partitions separately will reduce the num-ber of nearest neighbor searches and number of reachabil-ity distance computation. This model mine fraud transac-tions as outliers.
6 C
ONCLUSION AND
F
UTUREWORK
 
This paper presents the credit card fraud detection modelbased on clustering and outlier mining to find whether thecredit card transaction is fraudulent or not. Clustering algo-rithm is incorporated for the process of partitioning and out-lier detection algorithms are used to find the outliers. The out-liers are ranked and fraudulent cases are concluded. By apply-ing the algorithms in partitions separately will reduce thenumber of nearest neighbor searches and number of reacha-bility distance computation. The work can be further extendedby using density based Outlier Detection algorithm.
R
EFERENCES
 
[1]
 
M.J. Kim and T.S. Kim, “A Neural Classifier with Fraud Density Mapfor Effective Credit Card Fraud Detection,” Proc. International Confe-rence on Intelligent Data Engineering and Automated Learn ng, Lec-
 © 2010 Journal of ComputingPress, NY, USA, ISSN 2151-9617
http://sites.google.com/site/journalofcomputing/
m jiji
mm
1
1

Activity (2)

You've already reviewed this. Edit your review.
1 thousand reads
1 hundred reads

You're Reading a Free Preview

Download
scribd
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->