一种基于K-anonymity模型的数据隐私保护算法  

A DATA PRIVACY PROTECTION ALGORITHM BASED ON K-ANONYMITY MODEL

作　　者：谷汪峰[1] 饶若楠[1] 

机构地区：[1]上海交通大学计算机科学与工程系上海分布式实验中心,上海200030

出　　处：《计算机应用与软件》2008年第8期65-67,共3页Computer Applications and Software

基　　金：上海市科学技术委员会科研计划项目课题(05dz15005)

摘　　要：在数据共享的同时,如何保证数据的隐私性是一个重要的问题。泛化方法是数据隐私保护的一种重要方法,但现有的泛化算法不能处理连续属性,数据错误率比较高。在K-anonymity模型基础上,提出了一种扩展泛化算法EGA(Extended Generalization Algorithm),该算法在满足给定K值的条件下,用相对不具体的值最小限度地替换敏感数据,并实现了对离散属性和连续属性的处理。实验结果表明,与现有泛化算法相比,提出的算法具有运行效率高、数据错误率低、能保持敏感数据分类特性等优点。In the current information society, one important issue is the protection of data privacy during data sharing. Generalization is an important method of data privacy protection,but the existing generalization algorithms cannot process continuous attributes and produce a high data error ratio. The algorithm EGA （ Extended Generalization Algorithm） based on the K-anonymity model is proposed. Given a specific K value ,the algorithm partially replaces the sensitive data with data that are less specific as little as possible, and it can handle both discrete and continuous attributes. The experimental results show that compared with the existing generalization algorithms, EGA runs faster and produces a lower data error ratio. The classification character of the sensitive data can also be well preserved after generalization.

关 键 词：K-ANONYMITY 数据隐私 数据泛化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] TP18[自动化与计算机技术—计算机科学与技术]