Voronoi图划分实现位置数据发布隐私保护  

Voronoi Diagram Division for Location Data Privacy Protection Publishing

作　　者：薛佳楣[1] 张磊[1] 玄子玉[1] XUE Jiamei;ZHANG Lei;XUAN Ziyu(College of Information Science and Electronic Technology, Jiamusi University, Jiamusi, Heilongjiang 154007, China)

机构地区：[1]佳木斯大学信息电子技术学院,黑龙江佳木斯154007

出　　处：《计算机工程与应用》2019年第10期121-126,共6页Computer Engineering and Applications

基　　金：黑龙江省自然科学基金(No.F2015022);黑龙江省普通本科高等学校青年创新人才培养计划(No.UNPYSCT-2017149;No.UNPYSCT-2017175);佳木斯大学基础研究类项目(No.JMSUJCM2016-009);国家级大学生创新创业训练计划项目(No.201810222033)

摘　　要：针对位置这一特殊数据发布的隐私问题,提出了基于Voronoi图预划分的隐私保护策略。该策略通过信息熵计算处理待发布位置与敏感位置关联关系,并利用关联最低位置作为图心建立Voronoi图。进而利用Voronoi单元格特性将待发布的位置信息替换为图心位置,以此实现敏感信息隐藏的目的。在信息隐藏的基础上,利用广义差分隐私原理,提出了基于位置发布数据的ε-敏感位置关联隐私模型,并证明所提出的算法能够满足该模型。最后,通过比较实验进一步证明了所提出的算法在隐私保护能力和发布数据可用性方面的优势,并对实验结果进行了详细的成因分析。A privacy protection strategy based on Voronoi diagram division is proposed to solve the privacy problem of location. The relation between the position to be published and the sensitive position is processed by information entropy calculation, and the Voronoi diagram is established by using the lowest position of the relation as the graph center. Then,the position information to be published is replaced by the position of the graph center by using the Voronoi cell feature,so as to realize the purpose of hiding sensitive information. On the basis of information hiding and by using the principle of generalized differential privacy, a model for the sensitivity of location is proposed, and the results show that the proposed method satisfies the model. Finally, the comparative experiments further prove the advantages of the proposed method in the privacy protection ability and the availability of published data, and the experimental results are analyzed in detail.

关 键 词：信息技术 ε-敏感位置关联 VORONOI图 位置数据发布 信息熵 差分隐私 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]