数据发布中基于模糊集的隐私保护研究

Privacy-preserving in data publishing based on fuzzy sets

隐私保护数据发布是近年来研究的热点技术之一,主要研究如何在数据发布中避免敏感数据的泄露,又能保证数据发布的高效用性。基于模糊集的隐私保护模型,文中方法首先计算训练样本数据的先验概率,然后通过将单个敏感属性和两个相关联属性基于贝叶斯分类泛化实现隐私保护。通过实验验证基于模糊集的隐私保护模型(Fuzzy k-匿名)比经典隐私保护k-匿名模型具有更高的效率,隐私保护度高,数据可用性强。

Privacy-preserving data publishing becomes a hotspot technique in privacy preserving re- search, which mainly focuses on how to avoid leakage of sensitive data in data publishing, and at the meantime ensures efficient use of data. We propose a fuzzy k-anonymity model for privacy-preserving data publishing. We first calculate the prior probability of training sample data, and then achieve privacy protection through the Bayesian classification of a single sensitive attribute and two associated attrib- utes. Theoretical analysis and experimental results show that compared with the classic k-anonymity model, the proposal is more efficient, preserves more information, and has stronger practical applicability.