融合边信息的双重匿名位置隐私保护方案

Side Information Fusion Dual Anonymous Location Privacy Protection Scheme

数据挖掘技术的支持下,大量的服务数据、背景知识相互关联,所产生的用户查询偏好、行为模式等边信息会驱动攻击者实施更强大的位置隐私攻击手段,攻破了现有大多数基于k-匿名的保护方案.为了应对这些更强大的隐私威胁,本文提出一种融合边信息的双重匿名位置隐私保护方案SIFDA,即根据多样化查询概率和用户查询偏好相似性仔细筛选其他k-1个真实用户组成k-匿名集以抵御推理与共谋攻击.并设计了一个新颖的隐私保护度量标准,以精确衡量匿名集的隐私保护效果.同时,采用真实轨迹数据集,验证了本文所提出方法的有效性.

With the support of data mining technology,a large number of service data and background knowledge are related to each other,and the generated user query preference,behavior pattern and other side information will drive the attacker to implement more powerful location privacy attack means,which breaks through most existing protection schemes based on k-anonymity.In order to deal with these more powerful privacy threats,this paper proposes a side information fusion dual anonymous location privacy protection scheme SIFDA,that is,it carefully selects other k-1 real users to form k-anonymous set according to diverse query probability and user query preference similarity to resist inference and conspiracy attacks.A novel privacy protection metric is designed to accurately measure the privacy protection effect of anonymous sets.At the same time,the validity of the proposed method is verified by using the real trajectory data set.