Who Cares about Others' Privacy: Personalized Anonymization of Moving Object Trajectories

The preservation of privacy when publishing spatiotemporal traces of mobile humans is a field that is receiving growing attention. However, while more and more services offer personalized privacy options to their users, few trajectory anonymization algorithms are able to handle personalization effectively, without incurring unnecessary information distortion. In this paper, we study the problem of Personalized (K,∆)anonymity, which builds upon the model of (k,δ)-anonymity, while allowing users to have their own individual privacy and service quality requirements. First, we propose efficient modifications to state-of-the-art (k,δ)-anonymization algorithms by introducing a novel technique built upon users’ personalized privacy settings. This way, we avoid over-anonymization and we decrease information distortion. In addition, we utilize datasetaware trajectory segmentation in order to further reduce information distortion. We also study the novel problem of Bounded Personalized (Κ,∆)-anonymity, where the algorithm gets as input an upper bound the information distortion being accepted, and introduce a solution to this problem by editing the (k,δ) requirements of the highest demanding trajectories. Our extensive experimental study over real life trajectories shows the effectiveness of the proposed techniques.