The Research Of Community Structure Identification On Complex Network

In nature,complex network systems can be seen everywhere.Whether people can perceive economic,quotation and food chain network system,or people can not perceive the biochemical system,these systems have their own attributes and links.In order to fully study these systems,scholars abstracted a model for complex network.In recent years,the rising of complex networks has aroused the concern of experts in related fields,and also become the focus of research quickly.Through further research and analysis of complex networks,scholars find that different real network models have the same characteristics.Community structure is a key feature to describe the complex network,and it is also the most common and critical topology attribute in the network.The study of community structure not only has important theoretical significance,but also has practical application value.Community structure can help people better understand the network topology,function modules of complex network,and the hidden relationship between nodes,can also predict the trend of changes in the network system.In the community structure identification process,the modularity measure and its derived metrics play a very important role and facilitate the emergence of a large number of important community identification algorithms.The kind of general modularity optimization method to obtain the community structure has a resolution problem,affecting the modular optimization method of accuracy and breadth of the application.In order to solve the problem of resolution caused by optimization of modularity,this dissertation proposes an application of enhanced module degree optimization method,which can effectively avoid the resolution problem.As the division of the community structure and clustering algorithm similar to the idea,you can explore the use of data mining methods and theories to study the complex network of community structure problems.In this dissertation,the mature clustering algorithm is applied to the problem of complex network community identification.The contents of this dissertation are as follows:(1)The based on the enhancement module degree community identification algorithm: First of all,the algorithm uses the random walk theory to convert the undirected powerless network into the undirected right network.After the pretreatment of the network community between the edge of the small weight,thecommunity within the edge of the large weight.Then,the CNM algorithm is used to divide the actual network,and the modularity formula is used to measure the result of the partition.In this dissertation,a community identification algorithm combining random walk theory with CNM algorithm is proposed for the first time.The results show that this algorithm can effectively avoid the resolution problem caused by modular optimization.The algorithm is applied to the artificial network,the division of the community better.(2)An algorithm of community structure based on clustering algorithm: The algorithm is based on the information center degree of the edge given by Fortunato et al.,This dissertation proposes the node's intimacy and constructs the node affinity matrix.Then,the clustering idea is used to cluster the node affinity matrices,and a new algorithm of community structure discovery based on clustering is formed.As the clustering algorithm is sensitive to the initial value selection,this dissertation has developed some selection rules to effectively avoid such problems.Finally,the effectiveness of the algorithm is proved by the classical network model.