Credit Card Fraud Detection Using an Enhanced Artificial Neural Network with Regularization Techniques

Abstract

Credit card fraud detection remains a critical challenge due to the highly imbalanced nature of transaction data, evolving fraud patterns, and the need for real-time decision-making. Recent research demonstrates that ensemble learning techniques, such as bagging and boosting, significantly improve detection accuracy while reducing false positive rates by effectively handling class imbalance. In parallel, advanced deep learning architectures, including Convolution Neural Networks (CNNs) and Recurrent Neural Networks (RNNs), have shown strong potential in capturing complex spatial and temporal patterns within transaction data. Anomaly detection methods further enhance fraud detection capabilities by identifying irregular transaction behaviors and minimizing false negatives. This study highlights the importance of integrating ensemble learning, deep learning, and anomaly detection approaches to develop robust and adaptive fraud detection systems. Additionally, incorporating auxiliary data sources such as user behavior, geo-location information, and device fingerprints provides richer contextual insights, leading to improved prediction performance. The development of real-time fraud detection models is emphasized as a crucial research direction to prevent financial losses through timely intervention.