Social Roles Discovering Oriented Trajectory Pattern Mining Algorithm

With the popularity of smart phones,people's daily activities are recorded,forming large track data.How to analyze a large number of users' mobile trajectory data and mining the valuable information contained in these mass trajectory data is a hot topic in the current research of mobile objects.At present,the behavior pattern mining method does not take into account the semantic information of the residence time,residence interval and residence area of the moving object,and can not accurately excavate the characteristic behavior characteristic of the moving object.At the same time,the existing semantic pattern similarity measurement methods lack the consideration of spatiotemporal semantics and time randomness,and can not identify the behavior characteristics between the moving objects,thus effectively discovering the mobile object groups with similar social roles.In order to solve the above problem,this paper takes the feature of user behavior as the starting point,and aims at finding a mobile object group with similar social roles,and proposes a method of mining behavior pattern based on spatiotemporal Association.The specific work is as follows:Firstly,a semantic pattern mining algorithm based on spatio-temporal association is proposed.In view of the lack of three points of lack of considering spatial semantics,temporal semantics and regional temporal relationship,the improved Prefixspan frequent sequence pattern mining algorithm is the core technology,and the frequent semantic patterns of each moving object are excavated with the temporal and spatial semantics of the trajectory.The algorithm can accurately handle the time of the trajectory and the semantic level information of the residence point area,and effectively analyze and excavate the characteristics of the specific behavior of the moving objects.Second,a similarity measure algorithm between semantic models is proposed.The concept of temporal semantics is proposed to handle the semantic level information of stay time accurately and accurately.The concept of Spatial-temporalassociated semantics(STAS)is proposed to explain the rule of semantic similarity of trajectories.The concept of temporal entropy is proposed to measure the time randomness of trajectories passing through the same type of regions.Based on STAS,temporal entropy and temporal semantics,the temporal semantic and spatial semantic similarity measure(trajectory semantic similarity,TSS)between semantic patterns is given to describe the temporal and spatial characteristics of the social roles of the moving objects in which the trajectory belongs.Finally,the hierarchical clustering method is used to cluster the moving objects and find out the group of moving objects with similar behavior characteristics.Through the verification experiment,it is proved that the semantic pattern mining algorithm proposed in this paper can handle the semantic trajectory accurately and analyze the characteristic behavior characteristics of the moving objects,and the semantic similarity measurement algorithm can accurately and comprehensively discover the mobile object groups with similar life style or behavior habit.