Published Paper Details:

ABSTRACT: Scientific progress in electronic commerce and communication networks led to credit cards becoming the most popular payment instrument supporting in-store and electronic purchasing transactions. The number of credit card payment fraud incidents has substantially escalated because of current transaction trends. The yearly financial losses from fraud damages credit card issuers to such an extent that they need an advanced system with a complete fraud detection framework. Electronic payments have become a leading factor that drives the increase in fraudulent transactions. The urgent need exists for new methods which can detect credit card transaction fraud before it happens. Our fundamental research develops a transaction fraud system which integrates machine learning elements with original feature modification strategies. Our method received validation through experiments conducted with two real-world public credit card transaction datasets where one of them contained actual fraud cases. Our system achieved the highest success rate compared to other competing methods when processing both datasets. Our proposed method proves highly effective based on the achieved results. The tool from our approach strengthens credit card issuer capabilities to recognize fraudulent transactions thus safeguarding their customers while simultaneously minimizing loss and regulatory expense. A protected transaction environment for credit cards can emerge from implementing the procedures we have recommended.