Study Of The User Network Construction And Evolution Analysis Based On Behavior Trajectory Data

Spatial trajectory data can reflect human daily activity information and activity rules.Behind a large number of user location and behavior trajectory data,there are rich regular information about users' individual behaviors and users' social behaviors.Through the in-depth analysis and mining of these information,not only the individual behavior rule can be found from the microscopic angle,but also the common characteristics of the group(community)can be found from the macro view.Therefore,the study of user social relations and evolution laws based on their Spatio-Temporal behavior logs is of great significance to personalized recommendation,recommendation of application services,detection of group event,analysis of group behavior and so on.The main work of this paper includes the following aspects:(1)A user network construction method based on behavior trajectory data is proposed.Representation learning is designed to represent the semantic information of a research object as a dense low dimensional real valued vector,and it is one of the effective ways to accurately describe the research object.Therefore,in this paper,the correlation model and method of association rules association rules are used to analyze interest areas and standardize data for spatial trajectory data.Then,By introducing the word representation method(Word2vec)in Natural Language Processing domain,the SpaUser2Vec model is proposed,which is quantified to the spatial individual behavior,and the relation network between the users is constructed by using the operation method between the vectors.(2)A user network evolution method based on behavior trajectory data is proposed.the trend of user relationship change is analyzed,and a trend clustering method based on hierarchical clustering is proposed to analyze the trend of the relationship between users in the relationship network.Then,on the analysis of user community evolution,the evolution of user community is analyzed by time slice division and Louvain community discovery algorithm,and the evaluation of community discovery results is carried out,and the main factors that affect the evolution of user community are found.(3)The empirical analysis is carried out.The method and model proposed in this paper are tested by using the individual space trajectory data of university students.The results show that the proposed method can effectively express the behavior characteristics of the students,can reflect the relationship between individuals from the micro angle,and can also reflect the impact of key events on community evolution from a macro perspective.Therefore,the study of social relationship analysis and mining based on spatial trajectory data has important theoretical significance and application value for individual behavior analysis and group event detection.