异构身份联盟中基于本地差分隐私的信任评估算法  

CREDIBILITY EVALUATION BASED ON LOCAL DIFFERENTIAL PRIVACY IN HETEROGENEOUS IDENTITY ALLIANCE SYSTEM

作　　者：毛一凡 赵涛 于鹏飞 高先周 杨如侠 Mao Yifan;Zhao Tao;Yu Pengfei;Gao Xianzhou;Yang Ruxia(Big Data Center of State Grid Corporation of China,Beijing 100031,China;Global Energy Interconnection Research Institute Co.,Ltd.,Nanjing 210003,Jiangsu,China;State Grid Key Laboratory of Information&Network Security,Nanjing 210003,Jiangsu,China)

机构地区：[1]国家电网有限公司大数据中心,北京100031 [2]全球能源互联网研究院有限公司,江苏南京210003 [3]信息网络安全国网重点实验室,江苏南京210003

出　　处：《计算机应用与软件》2023年第2期339-344,共6页Computer Applications and Software

基　　金：毛一凡,高工,主研领域:系统架构;赵涛,工程师;于鹏飞,高工;高先周,工程师;杨如侠,工程师。

摘　　要：面对海量异构的身份管理、跨网跨域信任服务和身份隐私保护需求,构建异构身份联盟体系是一种有效的解决方法。针对联盟体系内关于信任评估计算中所存在的隐私泄露问题展开详细的研究,并利用本地差分隐私技术设计具有隐私保护效果的信任评价算法,对其隐私信息进行有效的保护。从理论上证明了该算法满足本地差分隐私,并分析其估计结果的误差大小,进行的相关实验论证说明算法的可行性与有效性。Facing the needs of massive heterogeneous identity management,cross-domain trust services and identity privacy protection,it is a feasible solution to construct the heterogeneous identity alliance system.This paper conducted the detailed research on the problem of privacy leakage in credibility evaluation calculation in the alliance system,and designed the privacy-preserving credibility evaluation algorithms with local differential privacy to protect the privacy information effectively.It could be theoretically proved that this algorithm satisfied the local differential privacy.The error of its estimation results was analyzed.The feasibility and effectiveness of the algorithm were demonstrated by relevant experiments.

关 键 词：异构身份联盟 本地差分隐私 信任评价 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]