基于隐私保护的有效保持匿名算法研究  

Research on Effective Algorithm Based on Privacy-preserving

作　　者：吕品[1] 于文兵[2] 

机构地区：[1]武汉工程大学计算机科学与工程学院,武汉430073 [2]武汉工程大学理学院,武汉430073

出　　处：《武汉理工大学学报》2009年第5期91-94,共4页Journal of Wuhan University of Technology

基　　金：武汉工程大学青年科学基金(Q200906)

摘　　要：介绍了数据发布环境下隐私问题,已知某一包含敏感信息的数据库,如何使得攻击者无法获取隐藏其中的敏感信息,而合法用户又能利用此进行有效地统计。针对这2个相互冲突目标,文中利用攻击者对某一已知元组的先验知识和后验知识来定义隐私,并利用统计查询的估计误差作为实用性的定义,在研究FRAPP方法的基础上,提出了2步操作的匿名算法,并与FRAPP法作了对比论证和实验验证,证明此法的隐私保护的有效性和实用性。The paper introdtuces the privacy problem in data publishing, given a database instance containing sensitive information, it is important to obtain a view from the anonymous database instance, on one hand attackers cannot learn any sensitive information from the view, and on the other hand legitimate users can use it to compute useful statistics. These are conflicting goals. Definition of privacy of the paper relates attacker＇s prior belief for a given tuple, with the posterior belief for the same tuple, and definition of utility is based on the estimates of counting queries. The paper puts forward the two steps anonym algo- rithm according to the method of FRAPP, studies and analyses the private and utility about the algorithm.

关 键 词：隐私保持 实用性 统计查询 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]