A Clustering-Anonymity Approach for Trajectory Data Publishing Considering both Distance and Direction

Abstract

Trajectory data contains rich spatio-temporal information of moving objects. Directly publishing it for mining and analysis will result in severe privacy disclosure problems. Most existing clustering-anonymity methods cluster trajectories according to either distance- or direction-based similarities, leading to a high information loss. To bridge this gap, in this paper, we present a clustering-anonymity approach considering both these two types of similarities. As trajectories may not be synchronized, we first design a trajectory synchronization algorithm to synchronize them. Then, two similarity metrics between trajectories are quantitatively defined, followed by a comprehensive one. Furthermore, a clustering-anonymity algorithm for trajectory data publishing with privacy-preserving is proposed. It groups trajectories into clusters according to the comprehensive similarity metric. These clusters are finally anonymized. Experimental results show that our algorithm is effective in preserving privacy with low information loss.

To cite this article: H. Jiang and K. Hu, “A Clustering-Anonymity Approach for Trajectory Data Publishing Considering both Distance and Direction,” in CIT. Journal of Computing and Information Technology, vol. 29, no. 1, pp. 1–12, 2022, doi: 10.20532/cit.2021.1005276.