Enhancing Fraud Detection in Portable Wallet Payment Systems using Machine Learning: A Hybrid Approach

Abstract

The rapid growth of portable wallets has made financial transactions more convenient but fraud has also increased as a result. Due to the dynamic and complex nature of these threats, traditional fraud detection techniques frequently fall short. Fraudulent activities increase in tandem with digital transactions, requiring sophisticated and flexible detection techniques. This research focuses on enhancing fraud detection in portable wallet payment systems by developing a hybrid model using machine learning techniques. The proposed model in this paper integrates behavioural biometrics and contextual data to improve accuracy while minimizing false positives in real-time. This approach is optimized for mobile environments, ensuring both security and efficiency. The study also identifies existing research gaps, such as security concerns and computational efficiency, and provides a comprehensive evaluation of the model's performance against traditional methods.