This book explores deep learning as a next-generation approach to online payment fraud detection in the face of increasingly complex and adaptive threats. Traditional rule-based or shallow learning methods are no longer sufficient. Through ten focused chapters, this book tackles challenges such as behavioral modeling, spatiotemporal anomaly detection, class imbalance, behavior drift, and graph-based inference. It applies advanced neural architectures including LSTM, GRU, GANs, GNNs, and spatiotemporal transformers. With a problem-driven structure, each chapter links real-world fraud problems to tailored neural solutions, validated on large-scale transaction data. This book blends theory, practical design, and empirical rigor, offering researchers and practitioners a foundation for scalable, adaptive, and reliable fraud detection systems.

This book explores deep learning as a next-generation approach to online payment fraud detection in the face of increasingly complex and adaptive threats. Traditional rule-based or shallow learning methods are no longer sufficient. Through ten focused chapters, this book tackles challenges such as behavioral modeling, spatiotemporal anomaly detection, class imbalance, behavior drift, and graph-based inference. It applies advanced neural architectures including LSTM, GRU, GANs, GNNs, and spatiotemporal transformers. With a problem-driven structure, each chapter links real-world fraud problems to tailored neural solutions, validated on large-scale transaction data. This book blends theory, practical design, and empirical rigor, offering researchers and practitioners a foundation for scalable, adaptive, and reliable fraud detection systems.