Research On Non-overlapping Community Detection Algorithm In Complex Network

The real world has large-scale individuals,and the relationship between these individuals can be abstracted into complex networks.Complex networks are usually described by graphs.The nodes and edges of graphs describe the individuals in complex systems and the relationships between individuals.Community structure is the characteristic of complex network,which is manifested in the close connection of nodes in the community and the sparse connection of nodes between communities.According to whether the nodes belong to a single community or multiple communities,the community structure in complex networks can be divided into overlapping communities and non-overlapping communities.Discovering the community structure in complex networks is an important research content of complex networks.Label propagation algorithm is very suitable for mining large-scale network communities because of its simple steps and near linear time complexity.The algorithm adopts the strategy of fair treatment of nodes,including the introduction of randomness in the process of label propagation,which leads to the stability and efficiency of community division results.Aiming at the problems of label propagation algorithm,this paper proposes a node importance sensitive label propagation algorithm.This method introduces the leader node,and initializes the label for the node according to the nearest neighbor characteristics of the leader node.On this basis,the node propagation tags are traversed in descending order according to the importance of the nodes.Finally,experiments are carried out on several real-world and artificial data sets,and the experimental results verify the effectiveness of the method.The hierarchical community detection algorithm aiming at modularity optimization is very suitable for mining hierarchical communities in the network because it does not need prior knowledge and the result of community division is stable.In the process of mining hierarchical communities,this kind of algorithm places too much emphasis on the connection strength between communities and ignores the similarity between communities,which leads to the inaccurate results of community detection,and only takes the maximization of modularity as the optimization goal,and the convergence speed of the algorithm is also restricted.Aiming at the problems of hierarchical community detection algorithm aiming at modularity optimization,this paper proposes a hierarchical community discovery algorithm integrating label propagation and modularity.The algorithm introduces the initialization community strategy based on community similarity,uses the community similarity to guide the label propagation process,divides the network into several small communities,and then merges these small communities with modularity increment to speed up the convergence speed,and then divides the hierarchical communities through modularity increment on the basis of the initial community.Finally,experiments are carried out on several real-world and artificial data sets,and the experimental results verify the effectiveness of the method.