The Research Of Sampling Representative Users On Social Network Based On Degree And Representaive Degree

With the rapid development of online social networks,how to extract representative users from a large number of social network users has become a problem to be solved.A basic problem in social network analysis is to find a subset of users to represent the original social network.The subset of users is the representative user we want to obtain.In recent years,researchers have conducted in-depth research on social impact analysis,maximization of influence,and graph sampling issues,but little research has been done on the extraction of representative users.Based on the existing researches,this thesis proposes the research of sampling representative users on social network based on degree and representative degree for the deficiencies of previous studies.The algorithm mainly includes the following aspects:(1)The sampled user is divided into multiple attribute groups according to attributes,and the attribute group representative degree is calculated according to the node degree and the node representative degree.The degree of node is determined by the number of associations of this node with other nodes.The representative degree is in some local topologies and people with larger weights should have greater representation.Attribute group representation refers to the degree of representation of a subset of users for all users of the attribute group.Thus,the sample representative user problem is transformed into a sample user subset problem that has the highest representative degree for all attribute groups;(2)Using the representative degrees of all attribute groups to calculate the quality function with the largest representative degree,through the increment of the quality function to measure the degree of representation of each unsampled user for all attribute group users,select the user with the largest representative degree as a new representative user.We used two sets of real data to compare the six algorithms.The results show that,compared with the existing algorithms,the accuracy rate of the research of sampling representative users on social network based on degree and representative degree a 4.1%-44.6% improvement in the average of different experimental sets and the algorithm is more time-consuming than the previous algorithm.