The Division Of Dynamic Social Networks Community Structure

The division of social network community structure is an important research direction of social network analysis and it is a research hotspot in recent years.The traditional static social network community find methods can not meet the need of the current social network analysis.Therefore,with the development of technology,research has gradually turned to dynamic networks.Research on the discovery of dynamic social networks is helpful to reveal the organizational principles,topological structure and dynamic characteristics of real networks.It is of great significance.The main work of the dissertation is as follows:(1)In order to identify the important nodes in the network that take into account the degree of community center and the bridging capability between communities,D-importance(DI)is proposed.The algorithm combines the k-shell and structural hole recognition algorithms to achieve the purpose of finding important nodes.The simulation results show that the nodes obtained by this algorithm have a high degree of centrality and bridging.(2)Aiming at the shortcomings of the traditional static community partitioning algorithm with high time complexity and inability to identify the number of community,a static community partitioning algorithm(SHP)based on structure hole and proximity is proposed.The main idea of this algorithm is to use the DI algorithm to find the initial node first,then calculate the node that is most similar to the community where the initial node is located and perform local clustering in combination with the local module degree,so as to achieve the goal of community division.The simulation results show that the community structure divided by this algorithm has a higher accuracy and modularity.(3)Aiming at the deficiencies of the existing dynamic community find algorithm based on incremental analysis,which ignores network mutation,an incremental community clustering(ILC)algorithm based on link clustering is proposed.The algorithm consists of two processes(SHP and Local Structural Hole Proximity(LSHP)based on the local structural hole and proximity).The main idea of ILC algorithm is that the network uses the SHP algorithm to find the community structure at the initial moment.Then discuss the situation of the network at other moments.If there is a network mutation,use the SHP algorithm to obtain the community structure at that moment.Otherwise,firstly use the LSHP algorithm to find the community structure at that moment,and then judge the stability of the community structure obtained.If the community is stable,the network community partition ends at this moment,otherwise,the community structure at that moment is regained using the SHP algorithm.The simulation results show that the community structure divided by this algorithm has a higher modularity and NMI.