A Cluster Based Approach for Credit Card Fraud Detection System using Hmm with the Implementation of Big Data Technology

摘要

In today's framework, Credit card is used for online transaction and security of the usage of credit card is also a big issue. Here, the fraud detection is done in two Steps. First is training of transactions and second is detection of fraud in the incoming transactions. In the first Step, sequences of transactions are added in to the system. In the second step, the fraud is detected in the credit card using the cluster comparison of the transactions. In first stage, HMM (Hidden Markov Model) is used with the expenditure behavior of card holders and categorized as low, medium and high expenditure behavior. The probability for every transaction sequence is calculated and checked with the threshold value. If any deviation is encountered then it is declared as fraud. To analyse the transactions into clusters clustering is used with the implementation of Big Data technology.