Research And Application On Graph-clustering From Bidirected Networks

Graph is a common method of modeling entity relationship. Community structure ofnetwork graph is one of the most important topological properties of complex network inthe real world. Community structure of network graph has the characteristics of connectedclosely between internal node inside the community and interconnected loose betweencommunities. Mining cluster-community-structure of complex network graph is of greattheoretical significance and application value for us to analysis of topology structure,understand its function, found in the chart the implied mode, predict the behavior of thewhole network.In order to explore and discover the group structure of large-scale network,graph is considered as a real network theory model and many graph clustering methodshave already proposed. These methods have a wide range of application in fields of theInternet structure, social structure, biological protein structure and the mutual referenceliterature relationThis paper mainly expounds the background, significance, research situation and themain problems facing in the current research on the network graph clustering method,giving the general framework of research method of mining in the complex network groupstructure. Briefly compared and analyzed the main advantages and disadvantages of thetypical graph clustering methods on mining network group structure. Consider theconnected quantitative bidirected networks as the research object, this paper proposed aclustering method based on quantitative bidirected networks, and clustering analysis theactual BBS data. First of all, according to the requirements of the data interface,previously dealt with the BBS data, including the establishment of database, and grab andimport of the BBS data. Then mining BBS database, get the relationship between the userand realize that the community on the large-scale complex virtual network.Through the application and inspection on different kinds of real and simulation DataSet, this paper verified that the algorithm is effective and practical. And make acomparison with other similar algorithm, the experimental results demonstrate that our method is more effective in the data mining of virtual community.