Analysis On Micro-blog Group Network Structure And Core Users Identification

With the rapid development of social networks, the quantity of users ofMicro-blog is expanding. Micro-blog is gradually replacing the traditional socialmedia as a major platform for exchange and dissemination of information corner.With the enhancement of the relationship between the user’s connection, users existas groups in Micro-blog. Micro-blog groups can not only promote linkages betweenusers, but also can accelerate the dissemination of information. Therefore theanalysis of Micro-blog groups network structure can get the main mode ofcommunication and interaction between user in groups. Research to identify thecore user can be used to help Micro-blog marketing and monitor public opinion.This paper use social network analysis to analyze the network of a mico-bloggroup, introducing the method of build groups network and aspects of analysis thebasic characteristics of group network. And we give the method of identify the coreuser of groups form different angels, including network structure, user activityindex and information dissemination.In this paper, we take the HIT student group of sina weibo as experimentaldata, introducing the groups discovery algorithm. And we introduce commonperspectives of user features analysis in groups, including Micro-blog quantity, fansquantity, concerned quantity. We build a concerned network and analyze its basiccharacteristics from density of network, centrality, small groups, small world andcore-periphery structure models. Furthermore, we put forwards different methodsof identifying core users. We connect degree centrality and betweenness centrality,ranking the user by the composite score form network structure angel. Given themeasurement of user activity, we give users activity index by computing the Micro-blog quantity, fans quantity, concerned quantity weights. Finally, we introduce Gindex which is used to evaluate the contribution of academic scholars in co-authornetwork. And the number of Micro-blog are forwarded equal to the number ofpapers cited, then we rank the users by G value form the information propagationangle.