Based On The Clustering Of The Complex Network Of Associations Discovery Algorithm

Complex network,which is one of the most important models and methods to study complex systems,has attracted plenty of attentions from the scientific community. Researchers have discovered several statistical properties that many networks seem to have.A property that seems to be common to many networks is community structure, Detecting communities automatically in networks has practical significance to analyze the topology and functioning of complex networks and has invalid help to predict the behavior of dynamic complex networks.So it is widely used in the field of WWW, social networks,biological networks and so on.This dissertation is devoted to using clustering technique to identify community based on the topology of complex networks.Recent algorithms to find communities in networks are compared and a flow of detecting communities when using clustering technique to detect community is proposed.Similarity measure between vertices,to a great extent,impacts the results of clustering algorithm.In this paper,MVV algorithm, which converts all vertices in network into vectors in high dimensional space,is proposed.In this high dimensional space,traditional similarity measures,such as distance measure,similarity coefficient,are used to measure the similarity between vertices.More clustering algorithms can be used to detect community structure in networks based on the similarity measure defined in high dimensional space. Hierarchical clustering,K-means algorithms are used to find communities in real-word and computer-generated networks in high dimensional space.Results indicate that mapping vertices into high dimensional space can improve the ability of clustering algorithm to detect community structure in network.We compare the ability of different algorithms to find communities in network and also use fuzzy clustering algorithm to find vertex which lying boundary at communities.Traditional spectral algorithm is improved base on the vector representation of vertices in networks.Finally,the development directions of communities in complex networks are analyzed.