A Hybrid Approach Using Maximum Entropy and Bayesian Learning for Detecting Delinquency in Financial Industry

A Hybrid Approach Using Maximum Entropy and Bayesian Learning for Detecting Delinquency in Financial Industry

Dharminder Kumar (Department of Computer Science and Engineering, Guru Jambheshwar University of Science and Technology, Hisar, India) and Suman Arora (Department of Computer Science and Engineering, Hindu College of Engineering, Sonepat, India)
Copyright: © 2016 |Pages: 14
DOI: 10.4018/IJKBO.2016010105
OnDemand PDF Download:
List Price: $37.50
10% Discount:-$3.75


The use of credit card has increased tremendously in the past few years because of the boom in the economy which has also resulted in the increase in the credit card fraud cases. Various leading banks and software development companies worldwide are taking serious measures to deal with the gravity of this situation. This paper proposes a framework for credit card fraud detection that will detect frauds using maximum entropy according to the irregular behavior of the customers in various transactions of credit card. The comparative study of above approach with existing approaches is also addressed. Results show the feasibility and validity of each approach.
Article Preview


“A Fraud is defined as the receiving of financial advantage or causing loss by implicit or explicit deception; it is the mechanism through which the fraudster gains an illegitimate advantage or loss. Fraud occurs in various types of areas like Telecommunication, Insurance, Health Care, Money Laundering, Computer Intrusion, Credit Card etc. Online banking has been the fastest emerging Internet activity. Billions of dollars are lost annually due to credit card fraud. The tenth annual online fraud report by Cyber-Source shows that although the percentage loss of revenue has been a steady 1.4 percent of the online payments for the last three years (2006 to 2008), the actual amount has gone up due to growth in online sales. The estimated loss due to online fraud is $ 4 billion for 2008, an increase of 11percent on the 2007 loss of $ 3.6 billion (Leggatt, 2008). Keeping these issues in mind we will work out the detection of fraud in the online use of credit card. The fraud involves illegal use of the card or the card information without the knowledge of the owner and hence it is an act of criminal deception. Bolton & Hand (2001) categorize credit card fraud in two types: application and behavioral frauds. The application fraud is where fraudsters obtain new cards from issuing companies and illegitimately use false information or phishing someone’s card details online.

Behavioral fraud can be of four types: mail theft, stolen/lost card, counterfeit card and ‘card holder not present’ fraud. To “analyze the spending patterns on every card and to outline “any inconsistency with respect to the “normal” spending patterns is the only way to detect frauds. In the last few years several approaches for the detection of frauds concerning credit card have been proposed.

The foremost challenge” identified by most of them is that they require labeled data for both genuine as well as fraudulent transactions, to train the classifiers”. Also from these approaches, a bulk of transactions is flagged as fraudulent but actually they are genuine. A large amount of money and time is spent by bankers to check whether the transactions are fraudulent or not. In contrast, we propose a fraud detection system which uses maximum entropy, does not require fraud signatures and is able to detect fraud purely on the basis of spending behavior of the customer. Our approach is non-parametric in nature and also maintains one-to-one relationship between variables. It reduces the number of False Positives (FPs) i.e., the transactions identified as fraudulent by FDS but actually they were genuine. Domain independency is also achieved in our approach. Comparison with existing techniques is also discussed.

The structure of paper is as follows: A brief introduction of related works is given after introduction. Next Section is the proposed fraud detection system. Then Methodology of proposed approach described. Then implementation and results are covered. Conclude the whole research in last section

Complete Article List

Search this Journal:
Volume 13: 1 Issue (2023): Forthcoming, Available for Pre-Order
Volume 12: 4 Issues (2022): 3 Released, 1 Forthcoming
Volume 11: 4 Issues (2021)
Volume 10: 4 Issues (2020)
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing