A Novel Approach to Credit Card Fraud Detection: Optimizing Recurrent Neural Networks

Credit cards are a leading payment method for online and offline transactions due to advancements in electronic payment systems. However, this popularity has led to an increase in scams, resulting in substantial financial losses for businesses and consumers alike. To maintain trust in electronic payment systems, it is essential to effectively identify fraudulent transactions. Many studies have explored various algorithms for fraud detection, primarily relying on traditional machine learning models. This paper focuses on both machine learning and deep learning approaches to identify the most effective models for detecting credit card fraud. We propose an optimized recurrent neural network (RNN) for detecting fraudulent transactions. Our experiments on two public benchmark datasets demonstrate that the RNN model achieved nearly 100% accuracy, with a macro average precision of 0.83, recall of 0.93, and F1-score of 0.87 for the first dataset. For the second dataset, it attained an accuracy of 99%, with a macro average precision of 0.70, recall of 0.89, and F1-score of 0.76, outperforming traditional machine learning models. The findings of this research contribute to mitigating financial losses and enhancing the security of payment systems through an effective model for credit card fraud detection.