UPI FRAUD DETECTION BY USING MACHINE LEARNING

Abstract

The growing growth in digital payments, particularly through UPI, has exposed the channel to high levels of financial fraud, hence creating a need for smart fraud detection systems. This project deals with Artificial Intelligence to Pay Shield: A Smart UPI Fraud Detection System Using Machine Learning. This is in contrast to traditional rule-based fraud detection systems that use predefined patterns, whereas the system dynamically assesses live transaction data to identify anomalies, suspicious behaviour, and potential fraud attempts. Through supervised and unsupervised learning models, the system classifies transactions, predicts fraud risk and delivers instantaneous alerts to users and financial institutions. Once fraud activity is detected, an automated response is triggered which blocks the transaction or facilitates multi-factor authentication (MFA) so that unauthorized transactions cannot take place. Moreover, combining blockchain The system also uses behavioural biometrics and device fingerprinting for security purposes, stopping account takeovers and unauthorized access. Through AI-powered fraud detection, real-time alerts, and secure data management, this system enhances UPI transaction security significantly, reducing financial losses and fostering user confidence in digital payment frameworks.