Credit card fraud detection based on transaction behavior

The proliferation of the EMV (Europay-MasterCard-VISA) chip card design in the credit card business mostly resolved the problem posed by the old Magnetic stripe card technology. However, several papers are starting to question the design and implementation of the EMV. This paper is suggesting that a detection model must be available to capture the possible anomalous transactions - a fallback in case the technology will fail. Several classifiers were evaluated during the model creation however only the Random Tree and J48 yielded the highest accuracy value of 94.32% and 93.50% respectively. By thorough analysis of these two (2) classifiers, it shows that the J48 is more fit in understanding the transaction logs data.