首页 | 本学科首页   官方微博 | 高级检索  
     

隐私保护的信息熵模型及其度量方法
引用本文:彭长根,丁红发,朱义杰,田有亮,符祖峰.隐私保护的信息熵模型及其度量方法[J].软件学报,2016,27(8):1891-1903.
作者姓名:彭长根  丁红发  朱义杰  田有亮  符祖峰
作者单位:贵州省公共大数据重点实验室(贵州大学), 贵州 贵阳 550025;贵州大学 计算机科学与技术学院, 贵州 贵阳 550025;贵州大学 密码学与数据安全研究所, 贵州 贵阳 550025,贵州大学 理学院, 贵州 贵阳 550025;贵州财经大学 信息学院, 贵州 贵阳 550025,贵州大学 计算机科学与技术学院, 贵州 贵阳 550025;贵州大学 密码学与数据安全研究所, 贵州 贵阳 550025,贵州省公共大数据重点实验室(贵州大学), 贵州 贵阳 550025;贵州大学 计算机科学与技术学院, 贵州 贵阳 550025;贵州大学 密码学与数据安全研究所, 贵州 贵阳 550025,贵州大学 计算机科学与技术学院, 贵州 贵阳 550025
基金项目:国家自然科学基金(61262073,61363068);全国统计科研重点项目(2013LZ46);贵州省教育厅创新团队项目(2013-09)
摘    要:隐私的量化是隐私保护技术的重要支撑,信息熵作为信息的量化手段,自然可以用于解决隐私度量问题. 基于Shannon信息论的通信框架,提出了几种隐私保护信息熵模型,以解决隐私保护系统的相关度量问题,主要包括:隐私保护基本信息熵模型、含敌手攻击的隐私保护信息熵模型、带主观感受的信息熵模型和多隐私信源的隐私保护信息熵模型.在这些模型中,将信息拥有者假设为发送方,隐私谋取者假设为接收方,隐私的泄露渠道假设为通信信道;基于这样的假设,分别引入信息熵、平均互信息量、条件熵及条件互信息等来分别描述隐私保护系统信息源的隐私度量、隐私泄露度量、含背景知识的隐私度量及泄露度量;以此为基础,进一步提出了隐私保护方法的强度和敌手攻击能力的量化测评,为隐私泄露的量化风险评估提供了一种支撑;最后,针对位置隐私保护的应用场景,给出了具体的信息熵模型及隐私保护机制和攻击能力的度量及分析.所提出的模型和隐私量化方法,可以为隐私保护技术和隐私泄露风险分析与评估提供可行的理论基础.

关 键 词:隐私保护  通信模型  信息熵  隐私度量  风险评估
收稿时间:2016/1/15 0:00:00
修稿时间:2016/4/14 0:00:00

Information Entropy Models and Privacy Metrics Methods for Privacy Protection
PENG Chang-Gen,DING Hong-F,ZHU Yi-Jie,TIAN You-Liang and FU Zu-Feng.Information Entropy Models and Privacy Metrics Methods for Privacy Protection[J].Journal of Software,2016,27(8):1891-1903.
Authors:PENG Chang-Gen  DING Hong-F  ZHU Yi-Jie  TIAN You-Liang and FU Zu-Feng
Affiliation:Guizhou Provincial Key Laboratory of Public Big Data (Guizhou University), Guiyang 550025, China;College of Computer Science and Technology, Guizhou University, Guiyang 550025, China;Institute of Cryptography & Data Security, Guizhou University, Guiyang 550025, China,College of Science, Guizhou University, Guiyang 550025, China;School of Information, Guizhou University of Finance and Economics, Guiyang 550025, China,College of Computer Science and Technology, Guizhou University, Guiyang 550025, China;Institute of Cryptography & Data Security, Guizhou University, Guiyang 550025, China,Guizhou Provincial Key Laboratory of Public Big Data (Guizhou University), Guiyang 550025, China;College of Computer Science and Technology, Guizhou University, Guiyang 550025, China;Institute of Cryptography & Data Security, Guizhou University, Guiyang 550025, China and College of Computer Science and Technology, Guizhou University, Guiyang 550025, China
Abstract:The quantification of privacy plays an important role in the privacy protection. Information entropy as a quantitative method of information can be used to solve the problem of privacy measurement. In order to realize the privacy metrics, several models of privacy information entropy are proposed based on Shannon''s Information Theory. These models include the basic information entropy model of privacy protection, the information entropy model of privacy protection with adversary, the information entropy model of privacy protection with subjective feelings and multi-source information entropy model of privacy protection. In these models, the information owner is assumed to be the sender, privacy attacker is assumed as to be the recipient, and the privacy disclosure course can be regarded as a communication channel. Based on these assumptions, the entropy, mutual information, conditional entropy, and conditional mutual information are introduced to represent measurement of privacy, privacy disclosure, and privacy and disclosure with background knowledge for the privacy protection system. Furthermore, the quantitative evaluation of privacy protection strength and adversary ability is provided to support quantitative risk assessment for privacy disclosure. Finally, the specific information entropy model, measurement and analysis of privacy protection algorithms, and adversary ability are supplied for location privacy protection application. The proposed models and privacy metrics can be used as fundamental theory for the privacy protection technology and privacy disclosure risk assessment.
Keywords:privacy protection  communication model  information entropy  privacy metric  risk assessment
点击此处可从《软件学报》浏览原始摘要信息
点击此处可从《软件学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号