一种基于BIRCH改进模型的数据流K-匿名发布算法  

A K-ANONYMITY PUBLISHING ALGORITHM FOR DATA STREAMS BASED ON IMPROVED BIRCH MODEL

作　　者：王璞[1] 陆剑江[2] 赵雷[1] 杨季文[1,2] 

机构地区：[1]苏州大学计算机科学与技术学院,江苏苏州215006 [2]苏州大学网络中心,江苏苏州215006

出　　处：《计算机应用与软件》2011年第6期282-285,共4页Computer Applications and Software

摘　　要：针对数据流持续、实时等特征,提出了一种基于BIRCH层次聚类的K-匿名隐私保护发布算法,改进了原BIRCH聚类模型,对准标识符中不同类型的属性进行同一度量映射,由聚类特征可加性合并了CF树中的相关子簇,控制了单个元组的最大发布时延。实验表明该方法具有良好的隐私保护效果和信息利用水平。In this paper we propose a k-anonymity privacy protection publishing algorithm based on hierarchical clustering technology of BIRCH for the characteristics of persistence and real-time of data streams,and make some improvements on original BIRCH clustering model.Same metric mapping is used on different attributes in quasi-identifiers and the corresponding sub-clusters in CF-Tree are merged via the additivity of the clustering features,thus the maximum release delay of a single tuple is in control.Experiments show that this algorithm has good effect on privacy protection and information utility.

关 键 词：K-匿名 BIRCH 数据流 数据发布 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]