Comparative Analysis on Fraud Detection in Credit Card Transaction Using Different Machine Learning Algorithms

Abstract

Fraud detection in credit card transaction is an important area of study which attains its importance in the increasing fraud risk posed on credit card by fraudsters to steal money from others’ accounts. In the last several years, to identify fraud in credit card transaction the applications of machine learning has shown optimistic result; however, machine learning in the credit card transaction also has certain challenges. There are numerous proposed machine learning algorithms that can be used to detect fraud cases; however, not every model has been proved to be the best since each technique works well with a certain data set, and fraudsters improve themselves to avoid detection by current fraud detection systems. This paper presents a comparative analysis of the credit card fraud detection methods based on supervised and unsupervised learning. Precision, recall, and F1-Score were used to evaluate each performance of all the models, as well as a ROC curve and AUC score for comparison. Since we compared nearly five different models, this report can be utilized as a research tool.