Abstract

Fraud detection is crucial for maintaining the security and integrity of e-commerce operations. Recent efforts have derived fruitful results in fraud detection research, ranging from conventional machine learning to deep neural networks to more recent graph-based methods and hybrid models that combine the above techniques. This study presents a comprehensive review of common fraud types and a critical analysis of state-of-the-art fraud detection techniques in e-commerce, followed by a discussion of open challenges and future directions.
This survey could be an ideal starting point for novice readers to gain a holistic understanding of the key concepts and techniques related to e-commerce fraud detection. It also serves as a reference for relevant researchers and practitioners to refresh their knowledge in the vibrant field of e-commerce fraud detection in face of evolving fraudulent threats.

Key words: E-commerce, Machine Learning, Fraud Detection, Deep Learning, Future Directions.