ELCCFD: An Efficient and Enhanced Credit Card Fraud Detection using Enhanced Deep Learning Principle

Credit card fraud poses a serious threat to financial institutions and their customers; hence, stringent detection protocols are necessary. This study introduces an approach known as Enhanced Learning for Credit Card Fraud Detection (ELCCFD) to enhance the accuracy of credit card fraud detection. To improve the fraud detection process, the proposed method combines the strengths of Convolutional Neural Networks (CNNs), AlexNet architecture, and Gradient Boosting Machines (GBM). The proposed approach begins with cleaning up the credit card data to get useful features, then trains a Convolutional Neural Network (CNN) using AlexNet to figure out complex patterns and representations on its own. This study generates a complete set of features by merging the CNN’s output with features generated using GBM. The final model is trained by using a combination of deep learning and other conventional machine learning techniques to achieve the best results. Experimental findings on benchmark datasets demonstrate the effectiveness of the ELCCFD methodology, achieving an accuracy rate of 98%. This study combines AlexNet with GBM to get a model to capture the complex patterns and is easier to understand with the feature importance analysis. With its strong accuracy and reliability, the proposed methodology offers a strong option to fight credit card fraud, and it shows the potential for actual use in financial systems.