A Method Of Community Discovery In Social Networks Based On Local Node Importance

The social network carries a large and complex information as people's lives migrate to the network,and gradually attracted the attention of scholars in related fields.Lots of research found that the community structure exist in the in the social networks,and the process of digging out of this community is called community structure discovery.On the basis of the results of community found on other social network-based technology research was able to depth.More and more community discovery methods have begun to emerge by years of research,but there is no one method get publicly recognized.In the paper,the traditional societies discovery method is to analyze some issues,such as complicated calculations,low stability and several other issues.Therefore,the overlapping improved label algorithm based on local important node is proposed.In the algorithm,the evaluation model of network node is firstly constructed by calculating local attributes,and the result of evaluation can be used to reduce the number of labels.Secondly,an “attraction”model is built to be used in the judge of multi-label nodes and the transformation of labels as propagating.Finally,the simulation experimental platform of community structure discovery algorithm is built.Through comparing with other algorithms,the effects of the algorithm's division and the stabilization of the algorithm are proved.The simulation experiment of this paper is implemented in Matlab.By running Several Synthetic networks and real networks on the experiment platform,the evaluation method of local important nodes and the improved label algorithm are verified that community could be discovered better,and the overlapping network community could be dug out effectively...