面向轨迹数据发布的优化抑制差分隐私保护研究

Research on Optimal Suppression Differential Privacy Protection for Trajectory Data Publishing

对待发布轨迹数据进行抑制处理可以有效降低用户隐私泄露的风险,在传统方法中,对数据集进行全局抑制处理使其满足LKC-隐私模型的方法降低了轨迹数据的可用性,因此,本文提出一种面向轨迹数据发布的优化抑制差分隐私保护算法(OSDP),该算法首先对轨迹数据集的最小违反序列中的点进行有效局部抑制判断,根据抑制优先得分决定抑制顺序,更新最小违反序列集,达到降低轨迹数据集中频繁序列损失率、轨迹序列损失率的目的.其次根据更新最小违反序列集后的轨迹数据敏感信息,建立分类树并向叶子节点中添加拉普拉斯噪声来提高待发布数据的安全性.实验表明,相较于其他算法,本文提出的OSDP算法降低用户隐私泄露风险的同时有效减少了数据损失率,提高了数据可用性.

The suppression processing of published trajectory data can effectively reduce the risk of user privacy disclosure.In traditional methods,the method of global suppression processing of data sets to satisfy the LKC-privacy model reduces the availability of trajectory data.Therefore,this paper proposes an optimized suppression differential privacy protection algorithm for trajectory data publishing.The algorithm firstly makes effective local suppression judgment on the points in the minimum violation sequence of the trajectory data set,determines the suppression order according to the rejection priority score,and then updates the minimum violation sequence set to achieve the purpose of reducing frequent sequence loss rate and sequence loss rate in the trajectory data set.Secondly,according to the sensitive information after updating the minimum violation sequence set,establishing a classification tree and adding noise to the leaf node to improve the security of the data to be published.Experiments show that the proposed algorithm reduces the data loss rate effectively,improves the data availability and reduces the risk of privacy disclosure.