Credit card fraud detection based on self-paced ensemble neural network

Abstract

Along with the significant increase in the number of credit cards, the number of credit card frauds worldwide is increasing day by day. At the same time, the development of Internet technology has led to the emergence of new fraud methods. The traditional credit card fraud detection methods can no longer meet the needs of the current credit card financial industry development. Identifying fraudulent credit card transactions effectively, quickly and accurately has become a major concern for banks. Methods combining expert rules and statistical analysis, decision tree methods, anomaly detection methods, and feature engineering methods are used in credit card fraud detection research. Among the many methods, deep learning is a new artificial intelligence method that has developed rapidly in recent years and is widely used in credit card fraud detection research. This paper uses a self-paced ensemble neural network (SP-ENN) model to learn credit card fraud transactions by dividing the datasets with different hardness, then identifying these transactions by neural networks, and finally performing a comprehensive evaluation. It was found that this model significantly outperforms other up-sampling or integration models in detecting credit card fraud data.