面向多敏感属性保护的p-覆盖k-匿名算法

A p-Cover k-Anonymity Algorithm for Protecting Multiple Sensitive Attributes

隐私保护数据发布是近年来数据挖掘研究中的一个热点.匿名是隐私保护数据发布的一种常用技术.针对当前大部分匿名方法只考虑单敏感属性保护的不足,提出一个p-覆盖k-匿名模型,用于具有逻辑依赖关系的多敏感属性保护,并基于该模型设计出一个支持多敏感属性保护的匿名算法kpCover.仿真实验表明,基于p-覆盖k-匿名模型的算法kpCover能有效解决多敏感属性的删除泄露问题,同时保证发布数据具有较高的数据质量.算法是有效可行的.

Privacy preserving data publishing is a hot topic in data mining research community. Anonymity is a popular technique in privacy preserving data publishing. Many anonymous methods have been presented for privacy preserving data publishing. However,most of the existing methods only consider protecting single sensitive attribute. In this paper,we firstly propose a p-cover k-anonymity model to protect multiple sensitive attributes with functional dependency. Then an optimal global-recoding algorithm based on this model is presented. The simulation experiments on real datasets show that the proposed model and algorithm can effectively solve the eliminate disclosure in multiple sensitive attributes, while ensuring high data quality of the released data.