Publication Details

Abstract

Credit card fraud detection remains a critical challenge in the financial sector, with fraudulent transactions becoming increasingly complex and difficult to detect using traditional rule-based systems. This project leverages Artificial Intelligence (AI) and Machine Learning (ML) techniques to develop a credit card fraud detection system capable of identifying suspicious transactions effectively and efficiently. The study focuses on two widely used supervised learning models — Logistic Regression and Random Forest — implemented using Python and the scikit-learn library. These models were trained and evaluated on publicly available transaction datasets, with careful attention to handling data imbalance, model evaluation, and feature selection. The system is designed to predict fraudulent activities with high accuracy while minimizing false positives, which are critical for operational efficiency in real-world applications. The implementation supports model training, testing, and prediction workflows, and the final trained models are integrated into a deployable pipeline. This work demonstrates that with proper pre-processing and model tuning, traditional supervised learning algorithms like Logistic Regression and Random Forest can serve as robust and interpretable solutions for credit card fraud detection. The simplicity, scalability, and real-time applicability of these models make them well-suited for deployment in modern financial infrastructures.