Fraud Detection in Credit Card Using DataMining Techniques

Abstract

Credit card fraud is a wide-ranging term for theft and fraud committed using or involving a payment card, such as a credit card ordebit card, as a fraudulent source of funds in a transaction. The purpose may be to obtain goods without paying, or to obtain unauthorized funds from an account. The most accept mode is credit card for both online and offline in today world. It provide cashless shopping at every shop in all countries. So as credit card is becoming popular mode for online financial transactions, at the same time fraud associated with it are also rising. In this paper Hidden Markov Model (HMM) is used to model the sequence of operation in credit card transaction processing. HMM is trained using Baum-Welch algorithm with normal behaviour of cardholder. If an incoming credit card transaction is not accepted by the trained HMM with sufficiently high probability, it is considered to be fraudulent.