Personalized trajectory privacy preserving method based on graph partition

Abstract

A (s,λ)-overlap was proposed which based on the angle and location between two trajectories to construct the relationship between them, and also measure the weight between two trajectories by trajectory angle and distance, to-gether to construct the personal trajectory graph model. Then a greedy partition method to was proposed construct trajec-tory k-anonymization sets by finding the approximate optimal k trajectories. Experiment results on synthetic dataset show the effectiveness and reasonableness of proposed method.

Keywords

• privacy security;personalized; graph partition;(s
• λ)-overlap;trajectory angel