Genetic Programming with CNN Optimization for Financial Fraud Detection

Abstract

Financial fraud detection poses a critical challenge in the contemporary digital economy due to its potential to inflict substantial harm on individuals, businesses, and financial institutions. In this research, we introduce an innovative approach that combines Genetic Programming (GP) with Convolutional Neural Network (CNN) optimization to enhance the accuracy and efficiency of financial fraud detection systems. Genetic programming is leveraged to evolve and optimize the architecture of the CNN model, tailoring it to the unique patterns and features inherent in financial transaction data. The primary objective of this proposed method is to autonomously discover and adapt the optimal CNN structure for fraud detection, thereby reducing the need for manual feature engineering and improving the model's capacity to generalize across various fraud scenarios. We conduct extensive experiments on real-world financial datasets, comparing the performance of our approach with traditional methods and standalone CNN models. The results underscore the efficacy of the proposed method, underscoring its potential to offer robust and adaptive solutions for financial fraud detection.