Ensemble Approach for Credit Card Fraud Detection Using Champion-Challenger Analysis
Harshit Muhal, Gaurav Khatri, Gaurav Kumar Dhama, Deepika Bansal · 2022 International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES) · 2022
Credit cards have evolved as a way of transferring money through E-payment systems, as digital payments develop to smoothly help corporate transactions around the world. Credit card fraud, on the other side, has continued to be a huge global danger to financial firms, with various reports and figures revealing the scope of the problem. In this study, we used three distinct ensemble models to implement a champion-challenger framework and evaluated the best model amongst them. With Accuracy, Precision, Recall, and F1-Score of 99.86, 99.73, 99.99, 99.86, Challenger 1, which is an aggregation of Random Forest, Adaboost, and LSTM, was crowned the winner.