Research And Implementation Of Community Discovery Algorithm With Membership Function

The research of complex network is an important science that mixed with sociology,informatics and computer internet and so on.With the further development of complex network research,more and more common features in complex network was found and known,especially the feature of community structure in complex network,which can not only help people clealy understand and analysis the network funchtion,but also be the research basis of many other properties in complex networks.At the first time,this paper describes the research background and significance of complex network in detail,then introduces the theoretical basis of complex network and the current research status about the complex network community detection algorithm research at home and abroad,and then summarizes the definition of community strcture,and also introduces several kinds of typical community discovery algorithm.Because most of the current community detection algorithm considers one factors that influence the community divided,and the community structure is fuzzy.This paper defines a membership function which can compute a node's subordinate level to commity,and it's based on the definition of membership in fuzz mathematics and the community structure's basic properties.On that basis,This paper presents a kind of heuristic community discovery algorithm with membership function(MCDA)in this text.The algorithm take this membership function as a measure for community detection,by computing one node's membership to each community,we can find which community it belongs to.By combining fuzzy clustering's basic thought and membership functions,we can get a node's specific membership degree to each community by utilizing MCDA algorithm.By considering various influencing factors,the MCDA algorithm can get high accuracy,at the same time has lower time complexity.In addition,on the basis of the MCDA algorithm,in view of the complex network in real world is always changing,this paper also designs a dynamic algorithm.The algorithm make up the shortage of the MCDA algorithm in solving dynamic problems,By using dynamic membership function and the idea of the incremental dynamic community discovery algorithm,it implements the community partition for the dynamic network based on the community results of original network.Finally both the theory and experiment have proved that the algorithm has good effect and fast running efficiency.