The Research On Temporal Trajectory Data Protection Method Based On Differential Privacy

With the rapid development of mobile Internet in recent years,especially the rise of smartphones with location-based functions,services based on location information rapidly merge into all aspects of people's daily life.Such as commonly used navigation software,social software,take-away software,buy software,etc.,which are based on location-based application services.At the same time,the location data generated by these application services can also be provided to other research institutes for mining to gain useful information to support certain decisions.However,the privacy issue of data has always been a central issue that constrained the development of location-based services.As a relatively new privacy protection technology,compared with the traditional privacy protection technology,the differential privacy protection method is not affected by the background knowledge of the attacker.It can achieve the protection purpose of data privacy by adding noise to the data.Because the differential privacy protection model is based on a rigorous mathematical model,the research on the usability of data becomes an important research subject under the premise of satisfying the differential privacy protection model.In this paper,the trajectory data with timestamps protection is studied,and the temporal privacy protection method based on differential privacy is proposed to protect the privacy of trajectory data and improve the data availability.The prefix tree index structure is used to construct the temporal trajectory data.The timestamps of nodes on each branch prefix tree of temporal trajectory data are gradually increased.At the same time,a generalization of multi-level sub-structure prefix tree index structure is designed,which has generalization layer and non-generalization layer.The feature of generalization layer is that there is no constraint on the location point in the domain,but the timestamp in the domain is constrained.In order to improve the data availability,a prefix tree of noise temporal data is constructed by specifying the prefix tree height,and the empty nodes are pruned by using the decision threshold.In the data release phase,according to the consistency constraint among prefix tree nodes,the consistency processing algorithm is designed to deal with the consistency of the prefix tree of noise temporal data and the algorithm proposed in this paper is satisfied with the privacy requirement of the differential privacy protection model.Finally,the performance of the algorithm proposed in this paper is tested on two data sets,the relative error and the true positive rate are used to verify the availability of the published data.Experiments show that the data privacy protection algorithm based on the differential privacy designed in this paper has high availability while maintaining privacy.