An End-to-End Machine Learning System for Real-Time Credit Card Fraud Detection with Interactive Deployment

Abstract

Credit card fraud detection remains a persistent challenge in financial systems due to extremeclass imbalance, evolving fraud patterns, and the need for real-time decision-making. This studypresents a comprehensive end-to-end machine learning system for detecting fraudulenttransactions using supervised learning techniques. The approach integrates datapreprocessing, Synthetic Minority Oversampling Technique (SMOTE) for imbalance mitigation,and model development using Logistic Regression, Random Forest, and XGBoost classifiers.Experimental results indicate that the Random Forest classifier achieves superior performance,reaching approximately 99.9% accuracy while maintaining a strong balance between precisionand recall. The trained model is deployed through an interactive web interface, enablingreal-time fraud prediction with probability-based confidence outputs. This work demonstrates thepractical applicability of machine learning in financial fraud detection by bridging the gapbetween theoretical modeling and deployable systems, aligning with recent advancements inapplied fraud analytics