Research on Privacy-Preserving Sharing and Collaborative Computing of Transaction Data for Intelligent Risk Control

Abstract

This paper conducts a systematic study of the challenges in privacy-preserving sharing of transaction data faced by the financial intelligent risk control sector. It first analyzes how data silos lead to specific operational bottlenecks, including fraud detection models that "cannot see and cannot track", credit assessment that "cannot clearly perceive and cannot accurately evaluate", and group-wide risk management that "cannot fully observe and cannot effectively prevent". It also reveals the dual dilemma of traditional plaintext data-sharing models in terms of legal compliance and commercial trust. On this basis, the paper proposes an "One Core, Two Wings, Three Drivers" application framework for privacy-preserving collaborative computing designed to resolve these conflicts. This framework takes business value and joint risk prevention as its core objective, supported by two wings-technical integration and governance collaboration-and driven by three key elements: scenarios, compliance, and performance. The paper provides a systematic and actionable theoretical reference and practical solution for financial institutions to achieve secure data value circulation and collaborative intelligent analytics across institutions and scenarios, while strictly adhering to data security and privacy compliance requirements.