面向多敏感属性的隐私保护方法

Privacy Preserving Methods for Multiple Sensitive Attributes

针对现有的多敏感属性数据发布方法中存在的隐私泄露问题,在分析多维桶分组方法的基础上,基于分解的思想,提出一种新的数据发布模型(l1,l2,…,ld)-uniqueness,同时给出相应的匿名算法。该算法考虑了等价组中敏感属性值的分布问题,对各个敏感属性单独处理,打破了敏感属性间一一对应的关系,可以抵御背景知识攻击和相似性攻击。理论分析和实验证明,该算法可以有效防止隐私泄露,增强数据发布的安全性。

For the privacy leak problems of the existing multiple sensitive attributes data publishing methods, based on the multi- dimensional bucket grouping approach and the idea of decomposition, a new data publication model is defined, named（l1 ,l2 ,…, ld ） -uniqueness, and the corresponding anonymous algorithm is proposed. The algorithm considers the distribution of the sensitive attribute values in the group, processes each sensitive attribute independently, and it can withstand the background attacks and the similarity attacks. The theoretical analysis and experiments show that the new method can effectively prevent the loss of priva- cy, and enhance data security.