分布式数据挖掘的隐私保持  

Privacy Preserving in Distributed Data Mining

作　　者：李宁致[1] 徐兰芳[2] 

机构地区：[1]华中科技大学生物医学工程系,湖北430074 [2]华中科技大学计算机系,湖北430074

出　　处：《网络安全技术与应用》2005年第10期78-80,共3页Network Security Technology & Application

摘　　要：分布式数据挖掘模型假定数据源分布在多个站点上,而各站点在进行分布式数据挖掘的同时需要隐藏私有数据以便保持隐私。本文将多方计算与数据挖掘技术相结合,在两点积运算的基础上提出安全的两点积运算公式,并将其简化,使得分布式挖掘算法的效能与集中式挖掘一致或近似,而又确保分布于不同站点的数据保持隐私。Distributed data mining assumes that the data sources locate on the distributed sites,and each site holds some part attributes of the global traction.When all the sites team up and dig for informations,privacy preserving for the individual data must be needed.In this paper,we proposes an improved secure approach based on two dot metdx algorithm by integrating cooperative computing with data mining.The new two dox metrix algorithm can keep the efficiency of distributed mining consistent with centralized mining without the sites exposing their private data sources.

关 键 词：分布式 数据挖掘 隐私保持 分布式数据挖掘 隐私 数据挖掘模型 数据挖掘技术 运算公式 挖掘算法 站点 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]