Research On Moving Object Social Relationship Discovery Methods In Location-based Social Network

The combination between social platforms with location-based technology has promoted the formation and development of location-based social network.Intelligent devices integrated with GPS,wireless network,satellite positioning have provided a great convenience for users to proceed location-tag,location check-in and location sharing in social platforms.These data from location-based social network contain a large amount of information,which can provide data support for location-based service related applications(such as: user social relationship discovery,intelligent transportation system,tourism recommendation,crime path tracking etc.)and urgently need researchers carry on comprehensive analysis and effective calculation.This thesis takes location-based social network data as research object,takes user social relationship strength calculation and user social relationship dynamic variety characteristic extraction as research target,and researches the flexible and comprehensive theory and method about user social relationship mining.The main research works are listed as follows:(1)This thesis has deeply analyzed the spatiotemporal interaction of locationbased social networks,taken a full consideration of users' behavior characteristics and spatiotemporal interaction information,and put forward user social relationship strength calculation based on spatiotemporal interaction(USRSC).This method adjusts user's interaction weights in different context by building user spatiotemporal interaction context,which can more accurately calculate the social relationship strength between users and more comprehensively discover social relationship between users.(2)This thesis has deeply analyzed the variation characteristic of user's spatiotemporal interaction behavior in location-based social networks,taken a full consideration of variation conditions of the users' behavior characteristics and spatiotemporal interaction information,and put forward dynamic social relationship calculation based on user behavioral characteristics drifting(DSRC).This method uses time slice technique to accurately,comprehensively and real-time quantitatively calculate the dynamic social relationship strength between users and extracts the dynamic variation characteristics of user social relationship by building user behavioral characteristics drifting model,which can effectively alleviate the sparsity problems of location-based social network data.(3)In order to deeply research location-based social network data,expand and enrich the theories and methods of user social relationship mining domain,intensify the combination between user social relationship mining theories and methods,design and develop a spatiotemporal social relationship calculation prototype system on the basis of the research results,such as user social relationship strength calculation based on spatiotemporal interaction and dynamic social relationship calculation based on user behavioral characteristics drifting.