Deep Learning Based Credit Card Fraud Detection in Electronic Payment Platforms

Abstract

The rapid transformation of payment system using digital platform has offered several advantages like seamless transactions, convenient, and easy to use, however it is also triggered massive digital fraud through credit card. This credit card fraud is an online crime where cyber criminals used unauthorized credit card to carryout financial transaction. To solve this problem, the aim of this paper is deep leaning based credit card fraud detection in electronic payment platforms. This was achieved with using the data of European credit card users with a sample size of 550000 records, including normal and fraudulent transaction cases. The dataset were processed and pre-processed before applying to train hybrid deep learning model of convolutional neural network (CNN) and Long Short-Term Memory (LSTM) respectively. The model was validated through comparative analysis with other individual models like LSTM, CNN. Results achieved reported accuracy over 85% for all models, while the hybrid upon comparism reported 98% accuracy as the best. The model is recommended to companies managing financial transactions to facilitate real-time detection of credit card frauds. Future works can expand this study using dataset from other part of the works, as this work is limited to detect credit card fraud in the European continents only.