Experimental Evaluation Unified Payment Interface (UPI) Fraud Detection System Using Elevated Deep Learning Methodology

Unified Payment Interface (UPI), especially in India, has completely transformed how money transfers are made within bank accounts. However, UPI can also be exploited for committing frauds, which undermines user trust and system credibility. Early detection of UPI fraud has therefore become a priority for banks and technologists to ensure secure and seamless transactions. The Unified Payment Interface (UPI) has revolutionized digital transactions in India but has also become a target for fraud. This paper presents an elevated deep learning methodology (EDLM) combining ShuffleNet and Support Vector Machine (SVM) for detecting UPI fraud. The model leverages pointwise group convolutions and channel shuffling in ShuffleNet for efficient feature extraction, which is then classified by SVM. Results show the proposed method achieves a high accuracy of 95 % and computational efficiency compared to traditional models, thus ensuring secure and reliable UPI transactions.