Paper Key : IRJ************397
Author: Prof. Dhananjay Subhash Mane,Mr. Prasad Vilas Gaikwad
Date Published: 04 Jan 2025
Abstract
With the swift expansion of digital transactions, the Unified Payments Interface (UPI) has become a favored and convenient method for financial exchanges in todays world. Nevertheless, the growing dependence on digital platforms has also contributed to an uptick in fraudulent activities. This paper introduces a robust UPI fraud detection system that utilizes advanced machine learning techniques to bolster the security of digital transactions. The suggested system harnesses a wide array of features, including transaction patterns, user behavior, and device information, to develop a comprehensive fraud detection model. Machine learning algorithms, such as supervised learning classifiers and anomaly detection methods, are used to analyze historical transaction data and uncover patterns that indicate fraudulent activities.The model utilizes a labeled dataset comprising both legitimate and fraudulent transactions, enabling it to effectively differentiate between normal and suspicious activity. Keywords: Transaction, Payment, UPI, Attackers, Fraudulent, Hoaxers, Money, Dataset. Random forest; decision tree; logistic regression; machine learning; gradient boosting method; confusion matrix.