UPI Guard: Intelligent UPI Fraud Detection System

Abstract

The rapid growth of digital payment platforms has transformed financial transactions by enabling fast and convenient money transfers. Among these platforms, the Unified Payments Interface (UPI) has gained widespread adoption due to its simplicity and real-time transaction capability. However, the increasing reliance on UPI has also led to a surge in fraudulent activities such as phishing attacks, unauthorized transfers, and scam-based payment requests. Detecting these fraudulent activities in real time has become a critical challenge for financial systems.This paper proposes a machine learning-based UPI fraud detection system designed to identify suspicious transactions and prevent financial losses. The system evaluates transaction attributes including amount, time, and receiver information to estimate the probability of fraud. A classification model is trained to categorize transactions as legitimate or fraudulent based on learned behavioral patterns. To improve detection accuracy, rule-based validation is integrated alongside machine learning prediction. Suspicious receiver identifiers and unusually high transaction values are flagged as potential threats. Transactions identified as fraudulent are automatically blocked, and repeated fraudulent attempts lead to account suspension. The system is implemented using Python, Flask, and MySQL, with an administrative dashboard for monitoring fraud trends and high-risk users. The proposed solution enhances digital payment security through intelligent and real-time fraud detection.