Research On Overlapping Community Discovery Algorithm Based On Multi-objective Optimization

Complex networks can model real systems in various domains.The use of technology to analyze the community structure can obtain the laws and information contained in the network,which can be used as a baseline for prediction.Therefore,it is essential to study overlapping community detection.Aiming at the problem of overlapping spreading range in the existing key node discovery algorithm based on non-greed strategy,a key node discovery algorithm based on information entropy Entropy Update is proposed.Entropy Update uses iterative selection strategy.Firstly,the influence of all nodes in the network is calculated,and then the node with the greatest influence is selected as the key node.Then,the influence of the node within the influence range is updated,and the above steps are repeated until all key nodes are selected.The SIR model is used for propagation simulation in the experiment,and the results show that the key nodes selected by Entropy Update have higher propagation ability.The fuzzy clustering overlapping community detection algorithm usually needs given fuzzy threshold and the number of communities,which is sensitive to the initial parameters.To solve this problem,a fuzzy overlapping community detection algorithm Fuzzy Cluster based Multiobjective Evolutionary Algorithm(FCMOEA)is proposed.The algorithm adopts two-stage optimization idea.Firstly,the key nodes of network are selected.To find the best non-overlapping community division,the optimization of community center nodes is performed on KKM and RC.Then,to find the best overlapping community division,the maximum extended modularity and the number of overlapping nodes are optimized to determine the overlapping nodes in network.Experiments show that FCMOEA can obtain better overlapping community division.