Private Trajectory Data Publication for Trajectory Classification

Trajectory classification (TC), i.e., predicting the class labels of moving objects based on their trajectories and other features, has many important real-world applications. Private trajectory data publication is to anonymize trajectory data, which can be released to the public or third parties. In this paper, we study private trajectory publication for trajectory classification (PTPTC), which not only preserves the trajectory privacy, but also guarantees high TC accuracy. We propose a private trajectory data publishing framework for TC, which constructs an anonymous trajectory set for publication and use in data services to classify the anonymous trajectories. In order to build a “good” anonymous trajectory set (i.e., to guarantee a high TC accuracy), we propose two algorithms for constructing anonymous trajectory set, namely Anonymize-POI and Anonymize-FSP. Next, we employ Support Vector Machine (SVM) classifier to classify the anonymous trajectories. Finally, the experimental results show that our proposed algorithms not only preserve the trajectory privacy, but also guarantee a high TC accuracy.