Fraudulent Transaction Detection in Credit Card by Applying Ensemble Machine Learning techniques

Abstract

In the present day scenario, fraudulent activities associated with financial transactions, particularly while using credit card, are observed to be occurring in a fast rate. Hence, a fraud detection system involving various detection techniques is very much essential for the financial institutions to sustain the goodwill from the customers. Several fraud detection techniques have been proposed by researchers as well as practitioners with application of various algorithms to find the pattern of fraud. In this study, the application of various classification models are proposed by implementing machine learning techniques to find out the accuracy and other performance parameters to identify the fraudulent transaction. Classification algorithms such as K-Nearest Neighbor (K-NN), Extreme Learning Machine (ELM), Random Forest (RF), Multilayer Perceptron (MLP) and Bagging classifier have been implemented to critically assessed their performances and the performances are evaluated. We have proposed a predictive classification model by ensemble of five individual algorithms, as it provides a better predictive performance.