A Hybrid Machine Learning-Based Multi-Stage Framework for Credit Card Fraud Detection

Abstract

The rapid growth of digital transactions has significantly increased the risk of credit card fraud, leading to substantial financial losses for both consumers and financial institutions. Traditional fraud detection systems often struggle with high false-positive rates and inability to adapt to evolving fraud patterns. To address these challenges, this paper proposes a hybrid machine learning-based multi-stage framework for efficient detection of credit card anomalies and fraudulent activities. The proposed framework integrates multiple machine learning techniques across different stages, including data preprocessing, anomaly detection, and classification. In the initial stage, data cleaning and feature engineering are performed to enhance data quality and extract meaningful patterns. The second stage employs anomaly detection techniques such as Isolation Forest and clustering methods to identify suspicious transactions. In the final stage, supervised learning models such as Random Forest, Support Vector Machines, and Logistic Regression are used to classify transactions as legitimate or fraudulent. This hybrid approach leverages the strengths of both unsupervised and supervised learning, improving detection accuracy while reducing false alarms. Experimental results demonstrate that the proposed system achieves higher precision, recall, and F1-score compared to traditional single-model approaches. The framework is scalable, adaptive, and suitable for real-time fraud detection in modern financial systems.