基于事务型K-Anonymity的动态集值属性数据重发布隐私保护方法

Privacy Preserving in Re-Publication of Dynamic Set-Valued Data Based on Transactional K-Anonymity

研究了动态集值属性数据重发布中的隐私保护问题.真实的数据随时间的推移因插入、删除、修改等操作而产生动态变化.更新后数据的重发布将面临攻击者使用历史发布结果对敏感信息揭露的风险.提出了一种面向动态集值属性数据重发布的隐私保护模型,延续使用事务型k-anonymity原则保护记录间的不可区分性,并通过维持记录中敏感元素在更新过程中的多样性和连续性阻止其被揭露.结合局部重编码泛化和隐匿技术降低数据匿名产生的信息损失,进而提出了完整的重发布算法.通过在真实数据集上进行的实验和比较,研究结果表明提出的方法能有效阻止敏感信息的泄露,并降低发布结果的信息损失.

Privacy preserving for re-publication of dynamic set-valued data is studied in this paper.In practical applications,data always dynamically change due to inserting,deleting and modifying.The updated data should be re-published and the sensitive information of which will confront the risk of being exposed by adversary using the historical published result.A privacy preserving model is proposed in this paper to protect in-distinguishability of the records by continuing using transactional k-anonymity,and prevent sensitive elements from being exposed by maintaining the diversity and continuity of the sensitive elements in the updating process.A re-publication algorithm is also proposed to reduce information loss of the anonymous result by integrating local recoding generalization with suppression.Real-world datasets are used in the experiment,experimental results and evaluations demonstrate that our approach can effectively prevent privacy leakage and reduce the information loss of publishing result.