改进成分分析的差分隐私高维数据发布方法

DIFFERENTIAL PRIVACY HIGH-DIMENSIONAL DATA PUBLISHING METHOD BASED ON IMPROVED COMPONENT ANALYSIS

针对高维数据发布中“维度灾难”所导致发布结果可用性较差的问题,提出一种改进成分分析的差分隐私高维数据发布方法ICAHDP。ICAHDP通过引入属性重要度来优化PCA,利用优化算法对数据进行降维,减少时间和空间的开销。该算法在数据发布的过程中引入基于互信息的评价机制,确定最优的主成分个数。考虑到高维数据中可能存在多个敏感属性,ICAHDP引入敏感属性偏好,结合最优匹配理论,设计敏感属性分级保护策略来满足个性化的差分隐私保护策略。实验表明,ICAHDP不仅保证了发布数据的隐私性,而且很大程度地提升了数据的准确性和实用性。

Aiming at the problem of poor availability of publishing results caused by "the curse of dimensionality" in high-dimensional data publishing,we present a privacy preserving data publishing with improved component analysis method(ICAHDP).ICAHDP improved the principal component analysis(PCA) by employing the attribute importance,and reduced the dimension of the data with the improved PCA,which reduced the time and space cost.ICAHDP introduced the evaluation mechanism based on mutual-information into data release,which determined the optimal quantities.Considering the existence of multi-sensitive attributes in high-dimensional data,ICAHDP introduced the sensitivity preference,combined the optimal matching theory,and designed the sensitive attribute hierarchical protection strategy.Extensive experimental results demonstrate that ICAHDP not only guarantees the privacy of published dataset,but also significantly improves the accuracy and data utility.