A comparison of machine learning algorithms for credit card fraud detection

Abstract

With the increasing use of credit cards for online and offline transactions, the risk of fraudulent activities has also increased significantly. In this study, we propose a machine learning-based approach to predict credit card fraud. We used a public dataset with 284,807 transactions, of which 492 were fraudulent. We experimented with various machine learning algorithms such as k-nearest neighbor, random forests, and isolation forests to develop predictive models for credit card fraud. We also performed feature selection to identify the most important features that contribute to credit card fraud prediction. Our results suggest that the proposed machine learning approach can effectively detect fraudulent transactions and can be adopted by banks and financial institutions to reduce the risk of credit card fraud.