Authors: Dr. Dhanalakshmi S, B. Sasi Prabha
Abstract: In an era where financial transactions are increasingly digital, the threat of cyber fraud has become a growing concern for both institutions and individuals. With every swipe, click, or transfer, there's a risk that sensitive data could be exploited by attackers using sophisticated techniques. As fraudsters become smarter, our defenses must evolve too. This chapter presents a practical approach to fraud detection using MATLAB, focusing on a simple, transparent, and explainable rule-based system. Rather than relying on complex machine learning models that can act as "black boxes," this method uses intuitive rules based on transaction amount, time, and location to flag potentially fraudulent activity. The system is built with ease of implementation in mind, making it ideal for financial institutions looking for an interpretable starting point or a lightweight solution for early warning detection. The model is demonstrated on simulated transaction data, and its results are visualized clearly to show the difference between normal and suspicious behavior. By the end of this chapter, readers will not only understand how to build a basic fraud detection system in MATLAB, but also appreciate the importance of balancing technical rigor with real-world usability in cybersecurity efforts.
DOI: https://doi.org/10.5281/zenodo.19718898
