Saad M. Darwish, Samah EL-Naggar, Saleh M. Elkaffas
Abstract The rapid digitization of financial transactions has increased efficiency but also exposed systems to sophisticated fraud attempts, posing significant challenges to ensuring transaction security. Traditional fraud detection approaches, including rule-based systems and conventional machine learning models, struggle to adapt to evolving fraud patterns, resulting in high false-positive rates and limited scalability. State-of-the-art methods, while leveraging deep learning, face limitations such as computational overhead, lack of transparency, and vulnerability to adversarial attacks. This study explores the integration of lightweight blockchain technology and deep learning for robust fraud detection in financial transactions. Lightweight blockchain ensures transaction immutability, transparency, and tamper-proof data sharing across nodes, addressing trust and security challenges. Meanwhile, deep learning provides dynamic and adaptive detection capabilities, employing neural networks to identify anomalous patterns in complex datasets. By reducing the computational and storage demands of traditional blockchain systems, the lightweight approach facilitates real-time fraud detection in resource-constrained environments, such as mobile and IoT devices. Our model combines these components, ensuring data integrity through blockchain while enabling efficient pattern recognition via deep learning, creating a system capable of addressing scalability, energy efficiency, and adaptability. Preliminary experiments demonstrate the model's effectiveness in reducing false positives, enhancing detection rates, and achieving scalability without compromising security or performance, marking a significant step toward secure and efficient financial ecosystems.