Research On Community Discovery Algorithm For Large-scale Complex Networks

Many real-world relationships can be described as forms of networks.Such as social relations networks,information and communication networks,protein structure networks,thesis co-authorship networks.How to dig out the hidden information in these networks has become a hot topic of academic research.Community structure is an important topology in a complex network.Community discovery of complex networks can help people understand the internal structure,relationships,and network properties of the network.Theoretical research on such subjects as sociology,biomedicine,economics,and computational science has important impetus.With the rapid development of information technology,the size of nodes in the network is increasing and the structure is more complex.The existing algorithms are mainly based on community discovery of the entire network and require global traversal.For a large-scale complex network,the time required to traverse the network is high.Therefore,this topic proposes to adopt the sub-network structure as the starting point for the discovery of largescale complex network communities,and to improve the computational efficiency of the algorithm by simplifying the scale of the network.A community discovery algorithm based on ant algorithm and modularity optimization(AMCA)is proposed.First,this paper first solves all groups of complex networks as sub-network structures.Then,the obtained subnetwork structure is used as a new node to reconstruct the network structure of a large-scale complex network and reduce the size of the original network.Finally,a modular degree optimization algorithm based on multiple merging is proposed to solve the community.The main research work of this article includes the following three parts:1)An improved ant colony algorithm(IACO)is proposed to solve all the structure of the network.By solving all the structure of the group in advance,the scale of the complex network is reduced and the solution speed is slow because of the large number of nodes in the large-scale complex social network,at the same time,the independence of the data is guaranteed.2)A network structure reconstruction algorithm(NSRA)is proposed.All the structure of the proposed group is used as a new node to restructure the edge structure between the unity points,thus reducing the network size,balancing the network structure of the complex social network and reducing the algorithm load in the community discovery phase.3)A multiple merge module degree optimization algorithm(MMOA)is proposed to solve the community structure in the reconfigurable network.Finally,on a large number of large-scale complex network datasets,it is compared with three community discovery algorithms with good experimental results.The results verify that the proposed algorithm achieves certain results in terms of solution quality and solution efficiency.