UPI Based Financial Fraud Detection Using Deep Learning Approach

This research seeks to tackle the essential issue of fraud detection in view of the increasing challenges presented by the changing landscape of online transactions, particularly inside the Unified Payments Interface (UPI). The research deviates from the usual methods used in this field and adopts a fresh approach by using Recurrent Neural Network (RNN) as an advanced instrument for detecting fraudulent financial transactions. This break with tradition shows that people have now come to terms with the fact that traditional methods have their limits and that CNN and RNN have something special to convey. By applying the algorithm in UPI transaction dataset, we segregate the fraud and legitimate transaction. To evaluate the proposed approach, we applied the test data on confusion matrix and got the True Positive Rate (TPR) is 87.5%, and the False Positive Rate (FPR) is 13.4%.