Font Size: a A A

Social Relation Analysis System Based On Spatial-Temporal Data

Posted on:2020-08-11Degree:MasterType:Thesis
Country:ChinaCandidate:H WangFull Text:PDF
GTID:2428330599958579Subject:Computer technology
Abstract/Summary:PDF Full Text Request
In today's society,people are happy to share their feelings on social networks,and a large part of these posts contain location labels.The interaction between time and space in these social networks contains a lot of information that can represent the relationship between people,and the analysis and mining of these data are very valuable work.With the increasing use of mobile devices,it is more convenient to collect users' location,such as GPS,Wifi and mobile phone tower.In this paper,a social relationship analysis system based on spatial and temporal big data is proposed,which can analyze the relationship strength and friend probability of users based on the spatial and temporal data collected from location-based social networks(LBSNs).The key technologies in the system will also be introduced in detail in this paper.MARS is proposed in this paper,which can measure the relationship strength between two people according to their location,distance and time relationship.Since private location and public place have different location weights,and different time periods also have different time weights,we take these factors into account in the model.This article also defines the spatial similarity and temporal similarity according to the check-ins of people,the more the user's trajectory is similar,the more frequently the interaction between the user and explain the relationship between the user more closely,and then we proposed a prediction probability model of the friend relationship between users.In the part of model validation and experimental evaluation,we use four real spatiotemporal data sets to evaluate the proposed MARS and probability model in this paper,and analyze the experimental results.The experimental results prove that the proposed models are correct and better than the state-of-the-art method.
Keywords/Search Tags:spatial-temporal big data, spatial-temporal factors, MARS, probability models
PDF Full Text Request
Related items