科研之友
微信
新浪微博
Facebook
分享链接
摘要
Trajectory data often provides information that is well applicable to real-world scenarios such as traffic planning and location-based advertising. Individual trajectory information may disclose sensitive personal data, thus necessitating privacy protection methods. Current methods assume and utilize the same privacy requirements for all trajectories, which can impact their protection and data utilization efficiency. This paper proposes a privacy protection method based on divided time intervals which satisfy different privacy requirements. The method works by constructing a privacy requirement matrix and running trajectory pre-processing based on the different privacy requirements for trajectories in different time points and locations. It uses trajectories that satisfy the (1, delta)-constraint to construct an undirected trajectory graph. By finding the trajectory corresponding to the edge and vertex with the minimum weight, k(i)-anonymous sets are then constructed with trajectories which share the same or similar privacy requirements. The Manhattan distance can then be applied to calculate the space between trajectories distance, which narrows the gap between the theoretical privacy protection and the actual protective effects. Comparative experiments demonstrate that the proposed method outperforms other similar methods in regards to both privacy protection and data utilization.
