Research On Trajectory Interest Point Mining Based On Label Propagation And Privacy Protection

In the era of big data,location sensing technologies such as mobile communication and sensing devices digitize the geographic location of vehicles,and a large amount of trajectory data is generated.Location data contains the behavioral characteristics of vehicles.Analyzing and mining trajectory data can provide better travel services for people.Vehicle trajectory data is crucial for applications of traffic analysis,location-based social networks,vehicles networks,and other data mining.If location or trajectory data is directly used,it may lead to privacy leakage and even threaten the personal and property safety of users.A clustering mining algorithm applied to vehicle trajectory data is proposed in this article.A new privacy protection algorithm is proposed to address the issue of privacy leakage during data transmission.The main research content is as follows:(1)An improved label propagation based trajectory interest point mining algorithm is proposed for mining interest point information in vehicle trajectory data.Firstly,the original trajectory dataset is processed,and the algorithm uses density clustering algorithm to initialize and cluster the processed dataset to obtain the initial cluster node set.Secondly,multiple attribute functions are designed based on the characteristics of the cluster node set to determine the label of each cluster node.Finally,the initial cluster node set is clustered using a label propagation based on clustering algorithm to obtain a set of interest points.The spatial and velocity attributes of vehicle trajectory data are considered in the clustering process,and the multidimensional information of trajectory data is integrated into the mining process.This algorithm can improve the utilization of data and the accuracy of interest point mining(2)An interest point mining mechanism that incorporates anonymized privacy protection is proposed.This mechanism integrates anonymized privacy protection algorithms into the interest point mining process of trajectory data.Before mining vehicle trajectory data,k-anonymity,l-diversity,and t-similarity models are added,which can effectively protect the privacy information of vehicles from being leaked.(3)A differential privacy protection algorithm for interest points based on an improved index mechanism is proposed,mainly used to protect the privacy information of interest points.Semantic privacy and trajectory characteristics are combined to redesign the scoring function of the differential privacy index mechanism.The algorithm randomly outputs trajectory points with lower sensitivity in the region of interest,and deletes other trajectory point information.Finally privacy protected data is published,and this algorithm effectively protects the privacy information in the region of interest points.We build a Py Charm environment to research on trajectory interest point mining based on label propagation and privacy protection.The experiments showed that the clustering algorithm proposed in this article effectively improves the accuracy of clustering,and the privacy protection algorithm proposed effectively enhances data availability and privacy protection degree.