Federated Fraud Scoring Models for Privacy-Preserving Transaction Intelligence
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Global financial systems face escalating threats from sophisticated fraud schemes, necessitating collaborative transaction intelligence among banks, payment processors, and fintech platforms. However, stringent data protection regulations and jurisdictional privacy laws hinder centralized data sharing. This paper introduces a federated fraud scoring architecture that enables distributed model training and inference without transferring raw transaction data. Leveraging federated learning, secure aggregation protocols, and differential privacy mechanisms, the proposed framework preserves data sovereignty while facilitating collaborative fraud detection. Experimental evaluations on benchmark financial datasets demonstrate that the federated model achieves performance comparable to centralized approaches while ensuring strong privacy guarantees. This architecture supports regulatory-compliant intelligence sharing and scalable fraud prevention across cross-border financial ecosystems.
Fields of Study & Subjects
Privacy-Preserving Technologies in Data
Imbalanced Data Classification Techniques
Blockchain Technology Applications and Security
Publication details
- Journal / source
- International Journal of Artificial Intelligence Data Science and Machine Learning
- Volume
- 7
- Pages
- 346–351
- Publication date
- 1 January 2026
- Language
- EN