隐私保护的分布式决策树分类算法的研究  

Research on privacy preserving distributed decision-tree classification algorithm

作　　者：申艳光[1] 邵慧[1] 张永强[1] 

机构地区：[1]河北工程大学信息与电气工程学院,河北邯郸056038

出　　处：《计算机应用研究》2010年第8期3070-3072,共3页Application Research of Computers

基　　金：河北省教育厅科学研究计划项目(2009421);河北省自然科学基金资助项目(F2008000752)

摘　　要：针对分布式决策树构造过程中的隐私保护问题,引入安全多方计算方法设计了可以保护隐私的分布式C4.5决策树分类算法。该算法适用于数据集垂直分布和水平分布两种情况,同时提出了一种新的隐私保护程度的度量方法。实验结果证明设计的隐私保护分布式决策树分类算法不仅很好地保护了原始数据不泄露,同时保持了较高的分类精度。To solve the problem of privacy preserving of the distributed decision-tree building process,this paper introduced the secure multi-party computation method and designed an algorithm of privacy preserving distributed C4. 5 algorithm ,which was applicable to the vertically and horizontally distributed dataset,and also proposed a new computation method of the degree of privacy protection. Experimental results demonstrate that the privacy preserving algorithm can well protect the original data from revealing,and keep high classification correct accuracy.

关 键 词：分布式数据挖掘 隐私保护 安全多方计算 C4.5决策树算法 垂直分布 水平分布 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]