Abstract: Aiming at the problem that traditional trajectory data publishing does not consider users' privacy preferences in different places, the paper proposes a data publishing method based on differential privacy for trajectory privacy classification. In order to satisfy users' privacy protection needs for data of different sensitivities, setting dwell and hotspot attributes, different privacy levels are assigned according to users' privacy preferences. The density-based clustering algorithm divides the high-density trajectory points into the same cluster and introduces the standard deviation to segment the trajectories uniformly, reducing the spatiotemporal complexity of processing the trajectory data. Construct a prefix tree of noisy trajectory segments, assign a privacy budget based on the weights of the trajectory privacy levels and the tree height, and introduce a Markov chain to limit the size of the noise added to the data. Experimental results show that the method proposed in this paper effectively balances data availability and privacy.

Key words: privacy protection, differential privacy, data publishing, location services, privacy classification