A Review Paper of Credit Card Fraud Detection Using Federated Learning

In this paper, the use of Federated Learning (FL) techniques in Credit Card Fraud Detection has been provided. Consumers and financial organizations alike are becoming increasingly concerned about credit card fraud. Conventional fraud detection techniques often rely on centralized data repositories, raising security and privacy issues. A viable substitute that enables cooperative model training without exchanging raw transaction data is FL. Throughout the work done till now, privacy preserving and addressing class imbalance were the two most necessary tasks. Privacy-preserving helps to train the model on distributed datasets held by individual banks without sharing the sensitive customer information and through addressing imbalance between fraudulent and legitimate transactions we can ensure effective model performance. In some papers, other technologies including Machine Learning were also used to demonstrate the benefits of FL.