A Comprehensive Survey of Machine Learning Techniques for Credit Card Fraud Detection

International Journal of Computer Science and Technology | Vol. 12, No. 9, September 2010 | pp. 62–68
DOI: 10.46882/2010/IJCST/000103

Review Article

Title: A Comprehensive Survey of Machine Learning Techniques for Credit Card Fraud Detection

Names of Authors: Sanjay V. Patel¹, Mei Ling Chang²

Authors’ Affiliations: ¹Department of Data Analytics, Indian Institute of Technology, Mumbai, India; ²Department of Computer Science, National University of Singapore, Singapore

Abstract: Financial institutions suffer multi-billion-dollar losses annually due to sophisticated credit card fraud operations. Machine learning algorithms have emerged as crucial barriers, yet selecting optimal configurations remains challenging given extreme data imbalance issues, where fraudulent transactions account for less than 0.1% of total logs. This paper reviews over 50 prominent literature sources published between 2005 and 2010 focusing on supervised, unsupervised, and hybrid machine learning frameworks. We evaluate algorithms including Support Vector Machines (SVM), Random Forests, Artificial Neural Networks (ANN), and Hidden Markov Models (HMM) based on their detection rates and false-positive frequencies. The survey analyzes data preprocessing approaches, specifically Synthetic Minority Over-sampling Technique (SMOTE) and principal component analysis, highlighting their role in resolving data skewness. Our comparative matrix indicates that hybrid models joining unsupervised clustering with supervised classification yield the highest Area Under the Curve (AUC) values, averaging 0.94 ± 0.02. We also identify open challenges such as real-time processing constraints, concept drift where fraud patterns evolve dynamically, and privacy concerns during cross-bank data sharing. Finally, we map future research trajectories in deep network architectures.

Keywords: Fraud Detection, Credit Card Security, Machine Learning, Data Imbalance, SMOTE, Comparative Survey

Manuscript Timeline: Received June 02, 2010; Revised July 15, 2010; Accepted August 02, 2010; Published September 15, 2010.

Citation: Patel, S. V., & Chang, M. L. (2010). A Comprehensive Survey of Machine Learning Techniques for Credit Card Fraud Detection. International Journal of Computer Science and Technology, 12(9), 62–68.

Subscribe to read the full article: https://www.internationalscholarsjournals.org/subscribe-to-read