摘要
个性化信息服务提高了Web查询精度,但同时也带来数据隐私保护的问题.尤其在面向服务的架构(SOA)中,部署个性化应用时,如何解决隐私保护,这对于个性化服务是一个挑战.随着隐私安全成为微数据发布过程中越来越重要的问题,好的匿名化算法就显得尤为重要.论文总结了前人研究中考虑到准标识符对敏感属性影响的k-匿名算法,提出了直接通过匿名化数据计算准标识符对敏感属性效用的方法以及改进的效用矩阵,同时为了更好地衡量匿名化数据的信息损失,论文中提出了改进的归一确定性惩罚的评价指标,从匿名化数据隐私安全的角度进行分析,实现了改进L-diversity算法,即基于信息损失惩罚的满足L-diversity的算法.它是准标识符对不同敏感属性效用的、并具有较好隐私安全的改进算法.
Personalized information services offer a promising way to improve the accuracy of Web search,but they bring about additional requirements related to data privacy preservation.Nevertheless,current SOA usually have one of the main barriers for deploying personalized search applications,and how to do privacy-preserving personalization is a great challenge.Privacy becomes a more and more serious concern in service-oriented information search,so good algorithms are in need to be designed.In this paper,the authors considered the previous research on k-anonymity involving the influence of Quasi-identifier on sensitive attribute Bottom-up k-anonymity and present a method for calculating the influence of Quasi-identifier on sensitive attribute Bottom-up k-anonymity through microdata directly and improved utility matrix.To better evaluate the information loss of the anonymity data,the authors also present a quality metric,both the two major factors：data utility and privacy guarantee are well preserved,Improved Normalized Certainty Penalty（INCP）.To achieve better privacy protection,the authors present a method based on the utility of Quasi-identifier which is L-diversity satisfied.
出处
《计算机学报》
EI
CSCD
北大核心
2010年第8期1315-1323,共9页
Chinese Journal of Computers
基金
上海市高可信计算重点实验室开放课题(07dz22304)资助~~
