Centrality analysis of online social network big data

Node centrality and network centralization are usually applied to study the behaviour of individuals and group in social networks. As a typical online social network, BBS reply network is firstly constructed in this paper, in which identifications registered on BBS are considered as nodes and reply articles set up links between them. Then, degree centrality and betweenness centrality of nodes are investigated, and the correlation between degree and betweenness centrality is discussed. Statistics show that BBS reply networks have positive correlations of degrees and betweenness centrality. Finally, network centralization is especially analyzed. Experimental results show that the network centralization decreases sharply by removing central nodes, which implies that central nodes are inclined to have greater influence and more power in online social networks.