基于模糊集的隐私保护方法研究

在数据发布的隐私保护研究中,针对k-匿名方法的复杂性高、效率低及数据可用性差等问题,从基于模糊集的角度出发进行隐私保护的研究,重点是对数值型属性的处理,提出了基于模糊集的最大隶属度(MMD)算法.该算法对敏感数值型数据进行模糊化处理,把其变成语义型数据,结合隶属度一起发布以达到隐私保护的目的.并通过实验进行了验证,基于模糊集的隐私保护方法与k-匿名方法相比,具有更高的效率,且信息损失要远远小得多,发布数据的可用性更好.

Fuzzy-based methods for privacy preserving

This paper did research based on fuzzy sets to overcome the high complexity, low efficiency and poor data availability of k-anonymity in the research of privacy-preserving data publishing. It focused on the processing of numerical attributes, and proposed the maximal membership degree algorithm. It fuzzed sensitive numerical attributes to semantic data which was released combining with membership degree. Verified through experiments, compared with k-anonymity methods, the MMD has better efficiency, furthermore, its information losses will be far smaller than k-anonymity and the availability of released data is better.