论文部分内容阅读
现有的微数据发布隐私保护匿名模型均没有考虑敏感属性间的关联关系,不能抵制基于敏感属性间关系的攻击.为此,论文提出抵制敏感属性关联攻击的(l,m)-多样性模型,该模型要求匿名数据的每个等价类中,每维敏感属性上多样性至少为l,并且当某一敏感值从等价类中删除后,该等价类剩下的敏感值仍满足(l-1,m)-多样性.本文也提出了实现(l,m)-多样性的两个算法—BottomUp算法和TopDown算法.实验表明,所提出的算法均能实现面向多敏感属性的(l,m)-多样性模型,有效保护多敏感属性微数据的个体隐私.
None of the existing micro-data publishing privacy-preserving anonymous models do not consider the association between sensitive attributes and can not resist the attacks based on the relationship between sensitive attributes.Therefore, the paper proposes a (l, m) -diversity model that resists the attack of sensitive attributes , The model requires that every equivalent class of anonymous data should have a diversity of at least 1 on each sensitive attribute, and when a certain sensitive value is deleted from the equivalence class, the remaining sensitive values of the equivalence class still satisfy (l-1, m) -diversity.At the same time, two algorithms -BottomUp algorithm and TopDown algorithm which implement (l, m) -diversity are proposed in this paper.Experiments show that the proposed algorithm can achieve multi-sensitivity (l, m) - a diversity model that effectively protects the individual privacy of multi-sensitive attribute microdata.