Card Sentry: An Intelligent Machine Learning System for Real-Time Credit Card Fraud Detection

Thigulla Vishnu Vardhan³, Galeeb Chaithanya³

Rayachoti Babu Ajay³, Dungavath Jeevan Kumar Naik³

Scholar, Computer Science & Engineering, Sandip University, Nashik, Maharashtra, India

Dr. Pawan Bhaladhare

²Professor (Guide), Computer Science & Engineering, Sandip University, Nashik, Maharashtra, India

Abstract:With the exponential growth of digital payment systems and e-commerce platforms, credit card fraud has emerged as a critical threat to financial security worldwide. Traditional rule-based detection mechanisms fail to adapt to the evolving and sophisticated strategies employed by modern fraudsters. This paper presents Card Sentry, an intelligent machine learning–based system that enables real-time detection and classification of fraudulent credit card transactions. The proposed system leverages a comprehensive pipeline encompassing data preprocessing, class-imbalance handling using SMOTE, feature engineering, and the deployment of ensemble learning models including Random Forest and XGBoost. The trained model is integrated into a Flask-based web application that provides an intuitive user interface for real-time fraud prediction, secure user authentication, and transaction history management using MongoDB. Experimental evaluations on the publicly available Kaggle Credit Card Fraud Dataset demonstrate that the proposed system achieves an ROC-AUC score of 0.98, a precision of 95.6%, and a recall of 93.8%, significantly outperforming conventional baseline methods. The system addresses key challenges such as highly imbalanced datasets, real-time deployment, and interpretability, making it suitable for practical financial applications.Keywords: Credit Card Fraud Detection, Machine Learning, XGBoost, Random Forest, Flask, MongoDB, SMOTE, Real-Time Classification, Imbalanced Dataset