基于自组织聚类的多敏感属性数据发布算法  

Multi-sensitive Attribute Data Publishing Algorithm Based on Self-organizing Clustering

作　　者：钱榕 周娟 金鑫 QIAN Rong;ZHOU Juan;JIN Xin(Graduate Department,Beijing Electronics Science&Technology Institute,Beijing 100070,China;College of Computer,Xidian University,Xi′an 710071,China)

机构地区：[1]北京电子科技学院研究生部,北京100070 [2]西安电子科技大学计算机学院,陕西西安710071

出　　处：《济南大学学报（自然科学版）》2018年第6期463-468,共6页Journal of University of Jinan(Science and Technology)

基　　金：国家自然科学基金项目(61402021);国家档案局科技计划项目(2015-B-10)

摘　　要：为了提升发布后数据的安全性,解决其易受攻击造成敏感信息泄露的问题,提出了(l,x,w)多样性模型;该模型引入信息熵概念,通过约束等价组在敏感属性上的多样性及均匀性来实现对敏感属性的安全保护;同时基于该模型,提出了多敏感属性数据发布的基于信息熵的l多样性聚类(entropy based l-diversity clustering,EBLC)匿名算法,该算法基于聚类技术,依据非敏感属性对元组进行聚类,在同簇中依照其敏感属性生成满足(l,x,w)多样性的等价组,泛化所有等价组得到发布数据。对EBLC算法进行的仿真实验结果表明,该算法有较好的运行效率,同时在敏感属性数目以及多样性改变情况下的信息损失以及发布后数据的抗攻击性均能得到较好保证。To improve the security of post-release data and solve the problem of sensitive information leakage caused by re-identification attacks of post-release data,a model called(l,x,w)diversity was proposed.The concept of information entropy was introduced into the model,and the security protection of sensitive attributes was achieved by controlling the diversity and uniformity of the equivalent group in sensitive attributes.At the same time,on the basis of the(l,x,w)diversity model,an entropy based l-diversity clustering(EBLC)anonymity algorithm for multi-sensitive attribute data release was proposed.Through clustering technology,the dataset was clustered according to the non sensitive attributes,and the equivalence group satisfying the(l,x,w)diversity was found in the cluster according to its sensitive properties,and all the equivalence groups were generalized to get the published data.The experimental results of the EBLC algorithm show that the algorithm has better operating efficiency,when the number of sensitive attributes and the number of diversity change,the information loss and the anti-attack performance of the release data can be better guaranteed.

关 键 词：l-多样性 隐私保护 多敏感属性 聚类 数据安全 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]