Releasing trajectory data can provide strong support for government and commercial organizations to make urban planning and decision-making, but there is a serious risk of privacy disclosure.The AC_TFIDF algorithm based on the TF-IDF statistics is proposed on the basis of the existing differential privacy protection trajectory. The algorithm follows the definition of differential privacy and can dynamically determine the degree of generalization at different moments in the trajectory. In the generalization process, the clustering center is replaced by the nearest one to further improve the availability of the published data. The feasibility and validity of the proposed algorithm is proved by the analysis and validation on the real data set.

privacy preserving； trajectory data； data publication； differential privacy