Achieving Anatomization Constraints in Dynamic Datasets

[摘要] Anatomy is one of the well-known privacy preservation models that are proposed to address privacy violation issues in released datasets. Unfortunately, we found that Anatomy is often sufficient to address privacy violation issues in datasets that are focused on performing a time of data releases. Thus, if datasets are dynamic (i.e., the data is updated when new data becomes available) and they are independently released, Anatomy can be insucient. That is, released datasets are satised by Anatomy constraints, they still have privacy violation issues from data comparison attacks such as iFRCA, iMRCA, iMRcMLA, dFLCA, dMLCA, dMLcMRA, SVM, and partition changing such that they are presented in this work. To address these privacy violation issues in released datasets, a new privacy preservation model is proposed in this work. Furthermore, we show that the proposed model is higher secure in terms of privacy preservation than Anatomy with extensive experiments.

[学科分类] 医学（综合）

[关键词] Privacy preservation models;Anatomization models;Dynamic datasets;Sequential Datasets