Personalized Differential Privacy Preservation Method for Trajectory Based on Regional Density Analysis

Abstract

As a reliable privacy protection method, differential privacy has been widely used in trajectory data release. However, the current trajectory protection method based on differential privacy lacks the analysis of the privacy requirements of each user and allocates the same privacy budget to all trajectory location points of the user, which cannot provide personalized trajectory privacy protection according to user characteristics. Aiming at this problem, we propose a trajectory protection method based on area density analysis. Analyze the stay area of each user, calculate the stay point and reconstruct the trajectory set according to the time and distance thresholds. Using the minimum spanning tree clustering algorithm based on local density peaks to obtain the privacy-sensitive location points and active hotspot areas of the user’s trajectory. According to the designed privacy importance degree expression, calculate privacy score of each sensitive location point, and assign them the proper privacy budget value. The experimental comparison on real data sets shows that this trajectory privacy protection method can better reduce the privacy budget waste and improve the availability of data compared with the traditional differential privacy trajectory protection method.