Credit card fraud detection using machine learning techniques

Abstract

The extensive use of the internet is perpetually drifting businesses to incorporate their administrations in the online environment. As a result of the development of e-commerce websites, people and monetary corporations count on online administrations to carry out their transactions. The ever-expanding utilization of internet banking associated with vast variety of online transactions has led to an exponential increase in credit card frauds. The fraudsters can likewise utilize anything to in uence the systematic operation of the current fraud detection system (FDS). Therefore, we have taken up the challenge to upgrade the existing FDS with the most potential exactness. This research intends to develop an e cient FDS using machine learning (ML) techniques that are adaptive to consumer behavior changes and tends to diminish fraud manipulation, by distinguishing and ltering fraud in real-time. The ML techniques include Logistic Regression, Support Vector Machine, na ve Bayes, K-nearest neighbor, Random Forest, and Decision tree. According to this study, the Decision Tree classi er has emerged as the most useful algorithm among the wide range of various strategies.