Detection of Credit Card Fraud Detection Using HPO with Inception Based Deep Learning Model

The use of credit for together online and in-person acquisitions has skyrocketed in recent years because to developments in electronic commerce and communication networks. Credit card fraud, meanwhile, has been on the rise and is a major source of annual losses for banks and other financial institutions. To reduce these costs, accurate fraud detection algorithms are essential, but their development is complicated by the fact that most credit card databases are highly unbalanced. Because of their static of input vectors to output vectors, traditional machine learning procedures are inefficient when used to identify credit card fraud. As a result, they are unable to accommodate the ever-changing purchasing habits of credit card holders. With the advent of deep learning technology, it is now possible to automatically and correctly detect credit card fraud. This research study systematically enhanced the Inception-ResNet-v2 method and used it on real-time data. Using a convolutional neural network, this study presents a detection model that operates from beginning to end. This model incorporates the best features of Inception convolution and residual networks to solve the deep network's training difficulty and increase the network's breadth. Hunter-prey optimization (HPO) model is also used to select the best characteristics, further improving classification precision. Each model incorporates stratified K-fold cross-validation and a real-world dataset for detecting credit card fraud from European cardholders. As such, the suggested model is evaluated next to its contemporaries. According to the findings, the proposed model has a higher accuracy than its predecessors by a significant margin (96.57%).