一种结合联邦学习的多节点异常交易数据协同分析方法

By constructing a consortium blockchain architecture and federated learning framework based on the improved PBFT consensus algorithm, and integrating multiple nodes for collaborative analysis of abnormal transaction data, the problems of low efficiency and insufficient accuracy in existing technologies are solved. This enables efficient and accurate identification and tracing of abnormal transactions, while meeting privacy protection requirements.

CN121981832BActive Publication Date: 2026-07-17SHANGHAI CRIMINAL SCI TECH RES INST +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202610445176.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-04-07
Publication Date
2026-07-17
Estimated Expiration
2046-04-07

AI Technical Summary

Technical Problem

In existing technologies, the analysis of abnormal transaction data in the governance of telephone fraud relies on traditional manual processing and independent data analysis by a single institution, resulting in low efficiency and insufficient accuracy. Furthermore, due to data privacy protection requirements, transaction data cannot be shared and exchanged between different institutions.

Method used

A consortium blockchain architecture based on an improved PBFT consensus algorithm is constructed, integrating nodes such as financial institutions, payment platforms, and anti-fraud centers. Trusted data is collaboratively trained through a federated learning framework, and smart contracts are used to ensure the traceability of data access control and analysis processes. Dynamic threshold adjustment and multi-node cross-validation are combined to improve the accuracy of abnormal transaction identification.

Benefits of technology

It has achieved a several-thousand-fold increase in the efficiency of abnormal transaction identification, reduced the false negative rate to below 2%, met privacy protection requirements, and significantly improved the pertinence and regulatory effectiveness of anti-fraud work.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121981832B_ABST
    Figure CN121981832B_ABST
Patent Text Reader

Abstract

本发明提供了一种结合联邦学习的多节点异常交易数据协同分析方法，属于联邦学习技术领域，包括：通过搭建改进型PBFT共识算法的联盟链，部署联邦学习平台，各节点预处理交易数据并提取基础、行为及风险特征，协同训练异常交易识别模型；采用动态阈值调整与多节点交叉验证优化判定效果，智能合约嵌入溯源模块追踪资金流向。该方案实现数据不出本地，打破数据孤岛，兼顾隐私保护与监管合规，显著提升异常交易识别效率与精准度，适用于金融反诈、资金安全防控等场景。
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Nuclear power safety report examination method and system, electronic equipment and storage medium

    CN121327159A

  • Real-time anti-fraud transaction intelligent analysis system and method based on multi-chain cooperation

    CN121746070A