TITLE:
Enhancing Mobile Money Security: A Multi-Layered Fraud Detection System Using Machine Learning and Multi-Factor Authentication
AUTHORS:
Mohamed Yayah Bah
KEYWORDS:
Mobile Money Security, Fraud Detection, Machine Learning, Multi-Factor Authentication, Random Forest, Gradient Boosting, Financial Technology, PaySim Dataset, Transaction Security, Real-time Detection
JOURNAL NAME:
Journal of Software Engineering and Applications,
Vol.19 No.4,
April
2,
2026
ABSTRACT: Mobile money services have revolutionized financial inclusion in developing economies, yet they face significant security challenges from increasingly sophisticated fraud attacks. This paper presents a comprehensive multi-layered fraud detection system that integrates Multi-Factor Authentication (MFA) with advanced machine learning algorithms to enhance mobile money transaction security. The proposed system employs a three-layer architecture comprising preventive measures, real-time fraud detection, and intelligent decision-making components. We evaluated three machine learning models—Logistic Regression, Random Forest, and Gradient Boosting—using the PaySim dataset containing 908,213 transactions with 8213 fraud cases. The Random Forest classifier demonstrated superior performance with 99.95%accuracy, 96.77% precision, 93.18% recall, and an F1-score of 94.94%. The system architecture incorporates a five-layer design featuring MFA, ML-based fraud detection, and an automated decision engine. Functional testing across six critical modules validated the system’s reliability and effectiveness. Our results demonstrate that combining preventive authentication mechanisms with intelligent fraud detection significantly reduces false positives while maintaining high fraud detection rates, making it suitable for real-world deployment in mobile money platforms.