Research On Privacy Protection In Trajectory Data Publishing

The popularity of GPS-enabled devices(such as smart phones and tablets)and location-based social software has led to the production of large amounts of trajectory data.The publication of these data has opened up new directions for the analysis,research,and understanding of human behavior.At the same time,however,the problem of leakage of personal privacy through trajectory data has become increasingly serious.Because even after the user identification information is deleted,it is possible to disclose the identity of the individual whose mobile record is in the track,so it should be released through the privacy protection method.However,the existing trajectory data anonymous method provides user privacy,but the data availability is very low.In this paper,we propose two algorithms to overcome these shortcomings by using trajectory data protection algorithms based on graph models and trajectory data protection methods based on prior knowledge.For graph-based data-anonymous methods,the relationship between track privacy and data utility is balanced by constructing a personalized anonymous model.According to the trajectory privacy and dynamic requirements of data utility determined by the context of the user's environment,a personalized anonymous model is constructed to balance track privacy and data utilization.When the trajectory k-anonymity set is selected,the influence of the trajectory similarity and direction on the anonymity process is considered.At the same time,both the growth of the anonymous area and the availability of the trajectory data are taken into account.A trajectory angle cosine sum is used to evaluate the trajectory similarity and direction,and the trajectory is used.Distance builds an anonymous area.Finally,the trajectory map model is constructed and the selection problem of the trajectory K-anonymity set is transformed into the constrained minimum spanning tree problem in graph theory.For anonymized data based on prior knowledge,using positional distance and semantic similarity,as well as user-specified data availability requirements,the track data is anonymized by prior knowledge.Generalize the point of the sensitive position in the original trajectory,and retain a certain degree of semantic similarity with the original trajectory.At the same time,the distance between the location point of the anonymous trajectory and the original trajectory is reserved to meet the user's demand for data availability.Taking into account the user's dynamic and personalized requirements for track privacy and data utility in different application scenarios,three personalized anonymity models are proposed to select the optimal trajectory kanonymity set that satisfies the requirements.By analyzing the privacy level and data utility of the chosen trajectory k-anonymity set under different requirements,the effectiveness of the two personalized anonymity models is evaluated and compared with other scenarios.Compared with the existing work,although the algorithm brings more time overhead,the average similarity and the information loss rate of the kanonymous metric loci are used to indicate that the scheme in this chapter can provide the optimal trajectory for the user's needs k-anonymity set.