基于聚类的数据敏感属性匿名保护算法*

为了防止数据敏感属性的泄露,需要对数据敏感属性进行匿名保护。针对l-多样性模型当前已提出的算法大多是建立在概念层次结构的基础上,该方法会导致不必要的信息损失。为此,将基于属性泛化层次距离KACA算法中的距离度量方法与聚类结合,提出了一种基于聚类的数据敏感属性匿名保护算法。该算法按照l-多样性模型的要求对数据集进行聚类。实验结果表明,该算法既能对数据中的敏感属性值进行匿名保护,又能降低信息的损失程度。

Clustering-based algorithm for data sensitive attributes anonymous protection

In order to prevent the disclosure of data sensitive attributes, it requires preserving the anonymity of data sensitive attributes. The current algorithm that has proposed to meet l-diversity is mostly based on the hierarchy, which can lead to unnecessary information loss. For this reason, this paper proposed a clustering-based algorithm for data sensitive attributes anonymous protection, it adopted an improved distance measure method which was from achieving k-anonymity by clustering in attribute hierarchical structures and combined clustering together, the algorithm in accordance with the requirements of l-diversity model clustering of data sets. Experimental results show that the algorithm can not only protect anonymity of sensitive attri-butes in data set, but also reduce the extent of information losses.