Enhancing UPI Fraud Detection Accuracy Using Isolation Forest: A Novel Machine Learning Approach

The Unified Payments Interface (UPI) has revolutionized how digital transactions are carried out in India, providing a speed-efficient and secure means, very easy to use. Nonetheless, the rising utilization of UPI has further elevated malicious activities, for example, unauthorized transactions, spurious payment links, as well as money fraud. As a result, our paper explores the new approach through which an innovative algorithm, such as the Isolation Forest (IF) algorithm, has been used to combat it. It finds the unusual patterns within an imbalanced dataset without the support of labeled data. Our model has focused attention on analyzing transaction behaviors besides detecting anomalies in a given dataset. In this case, the remarkable accuracy recorded is as high as 98%, far more than common usage of techniques like Random Forest, Support Vector Machines (SVM), and Logistic Regression. Real-world datasets were used to test and under such testing, it turned highly effective for real-time fraud detection. This research indicates the potential of sophisticated machine learning methods to make digital payment systems more secure and, consequently, protect users from falling into financial fraud.