Enhanced Approach for Credit Card Fraud Detection with Updated Grasshopper Algorithm and Deep Neural Network

Abstract

Credit card fraud detection is one of the most important problems in the financial industry, and it requires powerful and efficient ways to detect fraudulent transactions and safeguard customers. The proposed approach to fraud detection in this paper is new and based on the Grasshopper Optimization Algorithm (GOA) and Deep Neural Network (DNN). The two-stage proposed methodology includes feature selection by proposed GOA, GA, and PSO to achieve the best feature set and training and using a DNN for classifying transactions. Experimental results show that Grasshopper + DNN outperforms other combinations as regards precision, recall, and F-measure with precision = 0.9322, recall = 0.92113, and F-measure = 0.92663194. These results show clear improvements in fraud detection at the cost of false positives and false negatives. The method simplifies the complexity and enhances the efficacy and effectiveness of fraud detection systems and is thus a useful tool for financial institutions. Directions for future studies include incorporating other advanced optimization methods and machine learning models in an effort to enhance the detection capability.