A Hybrid Deep Learning Ensemble Model for Credit Card Fraud Detection
Emmanuel Ileberi, Yanxia Sun
Abstract
The rising volume of online transactions has concurrently increased the incidence of credit card fraud, presenting severe challenges to financial institutions and consumers alike. Traditional fraud detection techniques often fail to adapt to the sophisticated and evolving strategies employed by fraudsters, resulting in significant financial losses. This study introduces a novel hybrid deep learning ensemble model to enhance credit card fraud detection. The proposed model integrates Convolutional Neural Networks (CNN), Long Short-Term Memory Networks (LSTM), and Transformers as base learners, with Extreme Gradient Boosting (XGBoost) serving as the meta-learner. The experimental results demonstrate that the hybrid ensemble significantly outperforms individual base learners and traditional methods, with a sensitivity of 0.961, specificity of 0.999, and area under the receiver operating characteristic curve (AUC-ROC) of 0.972 using the European Credit Card Dataset. This research demonstrates the effectiveness of hybrid deep learning techniques in addressing the complexities of credit card fraud detection, thereby contributing to the development of more secure and reliable financial systems.