Research On Community Detection Algorithm Of Complex Networks Based On Leading Degree And Following Degree

Analyzing the community structure can obtain important information in complex networks,help control the networks,and predict the evolution of the networks.In the beginning,researchers focused on undirected networks and non-overlapping communities.However,there are large number of directed networks in the real world and overlapping structures generally exist in complex networks,overlapping community detection and directed network community detection have gradually become the focus of community detection research.Overlapping community detection has always been one of the key points and difficulties in complex network research.The method of detecting overlapping communities combined with hierarchies will be more suitable for complex networks in the real world because of real networks generally have hierarchical structures.In this paper,based on the leading degree and the following degree of nodes,combined with the network hierarchy,a following degree tree overlapping community detection algorithm model is built.The model reorganizes the network into a tree structure and discovers overlapping nodes and overlapping communities by dividing the tree.Experiments show that the model can detect overlapping communities combined with hierarchical,and can find overlapping nodes that other algorithms cannot find.The value of the extended modularity is greater than the comparison algorithm,and the results of community division are closer to the actual community structure.Directed network community detection needs to consider the direction information of the network,which makes community detection more complicated and difficult than undirected community detection.The key to improving the algorithm of undirected networks to directed networks is to integrate the directionality into the algorithm model.In this paper,a calculation method of the direct leading degree is designed by using in degree and out degree of nodes,then a directed network community detection algorithm model based on the direct leading degree and following degree is built.The network is reconstructed by the model and the reorganized network is divided into communities.Experiments show that the model of the direct leading degree can simplify the original network.The model has achieved good detection results compared with other directed algorithms and undirected algorithms on the computer artificial network.The model performs good on sparse and dense networks,is suitable for networks with different community sizes and can accurately detect the community structure in the citation network.