ENHANCEDDETECTIONOFFRAUDINUNIFIEDPAYMENTSINTERFACE(UPI) TRANSACTIONS USING GRADIENT BOOSTING METHOD

The widespread use of the Unified Payments Interface (UPI), which allows for smooth real-time transactions, has significantly changed digital payment systems.However, this growth has also led to a surge in fraudulent activities. This study presents an advanced fraud detection model based on the Gradient Boosting algorithm, renowned for its superior classification performance on imbalanced datasets. The model leverages advanced feature engineering to extract transactional, behavioral, and temporal features from real-world UPI transaction data. The model achieves a high predicted accuracy of 98.4% with a precision of 97.8%, recall of 96.9%, and F1-score of 97.3% through meticulous hyperparameter optimization. These results outperform several baseline classifiers. The proposed scalable framework significantly enhances the security and trustworthiness of UPI-based digital payment systems.