面向统计应用的隐私保护发布

现有的隐私保护模型多为通用发布方法，基于信息损失率的评价方法在具体应用中并不能很好地反映数据的发布质量。面向统计应用提出了一种基于交互的序列表发布模型，结合用户查询需求基于QI相容类最小覆盖，在不割裂QI和敏感属性的基础上进行数据发布。实验结果表明，该模型对于大数据集的隐私保护发布具有很高的查准率。

Preserving privacy for statistical application in data publishing

Current privacy preservation models are almost general data publishing methods.Evaluation of information loss can not reflect the quality of published data well for some specific application.This paper proposes a sequence table publishing model based on user interactivity for statistical application.The model considers requirements of users and publishes data based on min-covers of consistent class without separating quasi identifiers and sensitive attributes.Experimental results show that the proposed model has a good query accuracy for mass data sets.