Towards Maximum Efficiency: Combining ELM with BA for Credit Card Fraud Detection

Credit cards have become indispensable in modern times, offering convenience, security, and financial benefits. As fraudsters continually evolve tactics to evade detection, security systems must enhance their capabilities to combat fraudulent activities effectively. This study focuses on leveraging machine learning techniques, specifically the Extreme Learning Machine (ELM), for credit card fraud detection. We propose an evolutionary approach using Bat Optimization (BA) and Particle Swarm Optimization (PSO) algorithms to optimize ELM parameters. The study conducts experimental evaluations on three benchmarked credit card datasets with various activation functions. Our findings reveal that ELM optimized with BA performs better than PSO in enhancing credit card fraud detection accuracy.