基于R树的k-匿名技术研究  

Research on k-anonymity technolgy based on R-tree

作　　者：张晓琳[1] 李猛[1] 李素伟[1] 汤彪[1] 褚燕华[1] 

机构地区：[1]内蒙古科技大学信息工程学院,内蒙古包头014010

出　　处：《内蒙古科技大学学报》2010年第4期355-359,共5页Journal of Inner Mongolia University of Science and Technology

基　　金：国家社会科学基金资助项目(07XTQ003);内蒙古自然科学基金重点资助项目(20080404Zd21);教育部春晖计划基金资助项目

摘　　要：现有的隐私保护算法都是针对静态数据集的匿名处理,但是频繁接触的数据集大多是动态的,针对动态数据集隐私保护算法的研究将是未来的热点.基于R树的k-匿名技术,适用于动态变化的数据集.在动态R树的创建过程中,针对已有二路分裂算法不能保证孩子节点的"相似"性,提出了基于k-means多路分裂算法的R树k-匿名技术,较好地解决了对于固定的M,随着k值的增加,孩子节点的相似性问题.实验结果表明,融入了k-means多路分裂算法的R树k-匿名技术,有效地提高了k-匿名的质量和匿名表的查询效率.Privacy protection algorithms are currently used for proceeding anonymous static data sets,while most of the frequently contacted data sets are dynamic.Therefore,research of privacy protection algorithm for dynamic date sets is a hot issue.k-anonymity technology based on R-tree is suitable for dynamical changing data sets.During the process of building R-tree,original two-splitting algorithm can not guarantee the similarity of the child node.In this paper,k-means multi-splitting algorithm was proposed to replace the original two-splitting algorithm,and the problem of similarity of child node with increasing the value of k under fixed M was solved.The results of experiment showed that by merging the k-means multi-split algorithm,R-tree has effectively improved the k-anonymous quality and the query efficiency of the anonymous forms.

关 键 词：R树 K-匿名 k-means多路分裂算法 动态数据 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]