UPI FRAUD TRANSACTION DETECTION USING MACHINE LEARNING

Abstract

With the rapid adoption of digital payment systems, Unified Payments Interface (UPI) has become a popular platform for instant money transfers in India. However, the increasing volume of transactions has also led to a rise in fraudulent activities, including phishing, account takeover, and unauthorized transactions. Detecting such frauds in real-time is critical to ensuring user trust and financial security. This study proposes a Machine Learning (ML)-based approach to identify and prevent UPI fraud transactions. The system leverages historical transaction data, including transaction amount, frequency, location, time, and device information, to extract meaningful features for fraud detection. Various supervised ML models, such as Random Forest, Logistic Regression, Support Vector Machines, and Gradient Boosting, are trained and evaluated for their effectiveness in classifying transactions as legitimate or fraudulent. Performance metrics such as accuracy, precision, recall, F1-score, and ROC-AUC are used to assess model efficiency. Experimental results demonstrate that ML-based models can detect fraudulent UPI transactions with high accuracy, providing a robust solution for proactive fraud prevention. The proposed system not only enhances the reliability of digital payments but also strengthens financial security and reduces potential monetary losses for users and banks.