Research And Implementation Of Algorithm For Complex Network Overlapping Community Structure Discovery

Complex networks can be used to describe transportation networks,social networks,neural networks,scientific research cooperation networks and Internet,etc..With the in-depth study of complex networks,researchers have found that community structures exist in many real networks and some of them interrelated and overlapping.Based on this,we study the algorithms of the complex network community division.Complex networks are composed of nodes and edges.The nodes in the network can represent different entities,some of which may belong to many communities;and edges represent the relationship between entities.We mainly do the following works On the basis of the diversity of node neighbors and the randomness of node selection:1)Proposed Individual Conformity Evolutionary Algorithm(ICEA).According to the conformity and variation of community nodes,different community division results generated according to choose different neighbor of any node,and find a better community structure by comparing the community division modularity.We judge whether the boundary nodes of the community are the overlapping nodes by using Louvain Method with Neighbor Voting of Chen Junyu et al..So as to complete the division of the complex network of overlapping communities.2)Proposed Gene Fragment Covering Algorithm(GFCA).The basic idea is the individual which consist of nodes neighbor,selecting a gene fragment of other individuals randomly and covering corresponding position of the current individual.If the modularity improved after coverage,then the individual is selected as the current individual,otherwise,the original individual is retained.Experiments show that the algorithm has certain advantages and practical usability in real networks.In order to verify the community division effectiveness of the proposed algorithm ICEA and GFCA,we select Zachary Karate club,Dolphin Social network and American College football these three real data sets which commonly used to verify complex network algorithm,and comparative analysis is made on the module,the number of communities and iterations in the division of the community between our algorithms and algorithm GN,fast algorithm FN,algorithm NKFCM and PFCM.Experiments show that our algorithms ICEA and GFCA are better than the classical algorithms in the run time and division results.