基于轨迹频率抑制的轨迹隐私保护方法

A Trajectory Privacy Protection Approach via Trajectory Frequency Suppression

移动终端和定位技术的快速发展带来了轨迹大数据.研究者通过挖掘和分析发布的轨迹数据集,可获得一些有价值的信息.攻击者也可利用所掌握的知识对发布的轨迹数据集进行推理分析,以较高的概率推断出用户的隐私信息.轨迹抑制是一类实现隐私保护的重要方法,然而轨迹抑制的点数越多会造成数据效用越低.因此,在满足用户隐私需求的情况下,如何选择合理的抑制点来提高匿名处理后的数据效用是数据发布中要解决的重要问题.针对以上问题,文中提出两种基于轨迹频率的方案对轨迹数据进行匿名处理.第一种方案是根据情况抑制整条有问题的轨迹数据或向有问题的轨迹数据集中添加假数据;第二种方案是采用特定的轨迹局部抑制法对数据进行抑制处理.实验表明相对于已有方案,在满足同等隐私需求的情况下,文中方案处理后的数据效用提升了近30%.

The rapid development of mobile terminals and positioning technologies forms big data of trajectories.Researchers can obtain some valuable information through mining and analyzing the released data sets of trajectories.Taking advantage of the knowledge which attackers gained and analyzing the released data sets,attackers can infer the identities and private information of users at a high probability and accuracy.Trajectory suppression is one kind of methods to achieve privacy protection.Nevertheless,trajectory suppression leads to the lower utility of data.Therefore,with the demand of guarantying users＇ privacy,it is the main problem needed to be solved that is how to improve the utility of anonymous data through choosing rational suppression points in data publishing.Aiming at this problem,we propose two methods based on frequency in trajectories publishing to improve the utility of anonymous data.The first one suppresses the whole defective trajectory or adds fake data according to the situation.The second one realizes the privacy protection through employing specific partial suppression.Experiments show that,with respect to existing schemes,our scheme increases the data utility by almost 30% under the situation of meeting the same privacy requirements.