Research On The Key Issues Of Trajectory Privacy In Location-based Service

With the development of data mining technology and Location Based Service(LBS), people’s focus has shifted from the single point position data to the trajectory data which is correlated with time and space. By apply data mining techniques to trajectory data, we can get needs of users accurately, on the other hand It may expose the user trajectory privacy。Trajectory privacy is an extension to position privacy At present, the existing research on the trajectory privacy usually convert lines to points,which will be operated by traditional location privacy protectionIn order to meet the needs of the trajectory of privacy, this paper summarizes the related theory of trajectory privacy protection and location privacy protection, takes account of the trajectory characteristics, then proposes a new algorithm of trajectory Privacy Preserving Based on improved K-means clustering. The main idea of this algorithm is that the algorithm first preprocess the trajectory, and then the divided trajectory into k class according to the K-means clustering method. Finally, get the K anonymous sets of each class respectively. By different way of their generalization, these anonymous sets can be a representation of a class, or a representation of a region. To some extent, It can protect the user’s trajectory privacy without destroying the data availability.In the clustering stage, the K-means clustering algorithm has been successfully used, which can be useful to application of location privacy technology in trajectory privacy protection. And in the preprocessing stage two kind of trajectory data complement method had been proposed: Based on the average slope method and based on similarity of trajectory method. In the stage of anonymous processing, this paper proposes a random projection algorithm based on trajectory reconstruction, which can achieve high dimensional trajectory information dimension reduction., It also plays a role in the protection of trajectory privacy. It also plays a role in the protection of trajectory privacy. Besides, this stage it also introduces other ideas, which are the release of trajectory feature and the generalization of trajectory. At last, the paper introduces the measurement method of trajectory privacy protection, and proposes to use the average distance between trajectory to measure the algorithm.