Credit Card Fraud Detection via Integrated Account and Transaction Submodules

Globally, credit card fraud is a prevalent dilemma. Credit card fraud detection is a classification problem where one aim is to classify legitimate and fraudulent transactions in an adaptive and an automated manner. This paper proposes to utilize a novel hybrid scheme that integrates two mechanisms: a universal model and a unique model. The universal model is a static mechanism that inspects transactions without regard to the cardholder’s history or any other related transaction. It does so by implementing rules that are obtained via analyzing the complete population. On the other hand, the unique model is a dynamic, behavioral scheme that establishes a separate profile for each respective cardholder. In doing so, the model can establish a specific and accurate system that judges said cardholder’s transactions. It was found that the integration of the two models greatly enhanced the performance of the overall system. The system is inherently capable of handling the class imbalance problem that is usually prevalent in credit card fraud classification. The proposed framework was implemented and tested on a typical dataset. The proposed framework exhibited superior performance when benchmarked with similar frameworks. It showed a very high fraud detection rate, high balanced classification rate, high Matthews’ correlation coefficient and a very minimal false alarm rate.