Online Payment Fraud Detection using Machine Learning
Author(s)
Abstract
Online payment systems have significantly transformed digital transactions, enabling fast and convenient financial activities worldwide. However, the rapid growth of digital payments has also increased fraudulent activities, resulting in financial losses for both institutions and customers. Traditional rule-based fraud detection systems often fail to detect evolving fraud patterns. This paper presents a machine learning-based approach for detecting fraudulent online transactions. The proposed system performs data preprocessing, feature engineering, and model training using Logistic Regression, Random Forest, and XGBoost algorithms. To address class imbalance issues, the Synthetic Minority Over-sampling Technique (SMOTE) is applied. The models are evaluated using accuracy, precision, recall, F1-score, confusion matrix, and ROC-AUC metrics. Experimental results show that XGBoost provides superior performance in identifying fraudulent transactions with improved detection rates and reduced false positives. The system offers an efficient and scalable fraud detection solution suitable for real-world financial applications.
Keywords