水平划分多决策表下基于相对粒度的隐私保护属性约简算法  

Privacy-preserving attribute reduction algorithm based on relative granularity over horizontally partitioned multi-decision tables

作　　者：叶明全[1] 伍长荣[2] 

机构地区：[1]皖南医学院计算机教研室,安徽芜湖241002 [2]安徽师范大学数学计算机学院,安徽芜湖241002

出　　处：《计算机应用研究》2010年第10期3701-3704,共4页Application Research of Computers

基　　金：安徽省高校省级自然科学研究资助项目(KJ2010B241)

摘　　要：为了解决分布式环境中多个参与方在不共享各自隐私数据的情况下完成全局属性约简计算的问题,提出了一种水平划分多决策表下基于相对粒度的隐私保护属性约简算法。该算法基于相对粒度约简理论实现了分布式环境下全局属性约简的求解,利用半可信第三方与安全多方基础协议,设计了安全多方计算相对粒度协议,使各参与方在不共享其隐私信息的前提下达到集中式属性约简的效果。分析结果表明,该算法是有效可行的。Aiming at the horizontally partitioned multi-decision tables,this paper proposed an algorithm based on relative granularity for privacy-preserving distributed attribute reduction,which could solve the problem that multiple parties carried out attribute reduction computation in distributed environment without sharing private data. The algorithm could compute global attribute reduction based on the attribute reduction idea of relative granularity,and used semi-trusted third party and secure multi-party technology to design a privacy-preserving protocol for computing global relative granularity,which could get accurate attribute reduction effect in the premise of no sharing of private information among participators. Analysis results show the proposed algorithm is effective and efficient.

关 键 词：分布式属性约简 相对粒度 隐私保护 安全多方计算 粗糙集 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]