Investigation into the Efficacy of Various Machine Learning Techniques for Mitigation in Credit Card Fraud Detection

Abstract

An effective credit card fraud detection model is the most challenging issue for the financial organizations. Statistical and machine learning (ML) techniques are widely explored in financial applications. But there is no thumb rule which technique gives better performance. Recent studies conclude that ensemble learning may be the right approach in this problem domain. In this paper, we aim to develop a novel fraud detection system using an ensemble model. In the proposed model, initially the imbalanced credit card dataset is balanced using random under-sampling technique, then the performance of the model is evaluated using both single base classifiers and ensemble of classifiers. In the proposed model, AdaBoost, random forest (RF), extreme gradient boosting and gradient boosting decision tree (GBDT) are used as ensemble models. The experimental result shows that the combination of RF and GBDT for the ensemble model is superior in performance as compared to other combinations.