Research On Community Detection Algorithm Based On Weighted Network

Community discovery is one of the commonly used techniques for mining complex network data.It reveals the community structure of the network,which is of great significance for in-depth understanding of-the network structure and mining of the deep information of the network.It is widely used in recommendation systems,biomedicine,public safety and other fields.Aiming at the problem that the existing static and dynamic community discovery algorithms cannot be fully applied to the discovery of mixed-form network communities,the paper studies the weighted network community discovery algorithm.The main research contents are as follows:A static weighted network community discovery algorithm based on mixing similarity matrix is proposed.The algorithm mainly includes two parts:the calculation of mixing similarity and the division of mixing similarity matrix clustering.Firstly,it defines the mixing similarity measurement method of balancing node attributes and edge weights;secondly,constructs the mixing similarity matrix of the network,decomposes and clusters it,and obtains the community structure of weighted network.The experimental results show that compared to the LPA algorithm,Louvain algorithm and IEM algorithm,the proposed algorithm has a higher partition quality on four data sets of different sizes;by analyzing the similarity of the difference nodes,it is found that the partition structure of the paper algorithm is more reasonable.The feasibility and effectiveness of the algorithm are verified.A dynamic weighted network community discovery algorithm based on adding node similarity is proposed.The algorithm uses an improved label propagation algorithm combined with node importance to divide the network at the initial moment;in the division of the subsequent networks,two types of adding nodes are defined in combination with node similarity,and different types of adding nodes are dynamically divided.The community structure of the dynamic weighted network is obtained.The experimental results show that compared to the DG algorithm and the IG algorithm,the algorithm has a higher modularity in the dynamic weighted real social network,and the division result of the algorithm is significantly more stable over time;In addition,the algorithm can describe the evolution trajectory of nodes and communities in dynamic networks.It proved the feasibility of the algorithm,effectiveness and stability.The MSWCD algorithm proposed in this paper solves the problem that the edge weight influence cannot be measured in the node similarity calculation method.The ADWCD algorithm can find the continuous community structure in the dynamic weighted network,and find the evolution events of the network.The experiment fully proves the effectiveness of the algorithm.Finally,the direction of community discovery algorithm worthy of further research and exploration is proposed.