A Secured Approach to Credit Card Fraud Detection Using Hidden Markov Model

As the usage of credit card has increased the credit card fraud has also increased dramatically. Existing fraud detection techniques are not capable to detect fraud at the time when transaction is in progress. Improvement in existing fraud detection is necessary. In this paper Hidden Markov model is used to detect the fraud when transaction is in progress. Here is shown that hidden markov model is used to detect credit card fraud with reduced false positive transaction. HMM categorizes customers profile as low, medium and high and based on spending profile set of probability for amount of transaction is assigned to each cardholder. Amount of new transaction is checked against the profile of card holder, if it justifies a predefined threshold value then transaction is accepted else it is considered fraudulent. But still HMM is not secured for initial some transaction during training so HOTP is used as secured approach with HMM to reduce the fraud and to increase the security. HOTP is one time password which is used one once and it is send to the client mobile when HMM detects that amount is more than threshold value if the user enters valid HOTP then only transaction is allowed to progress else it is detected as fraud.