Recursive Community Detection In Complex Networks

With the in-depth study of physical properties and mathematical characteristics of complex network properties,the researchers found that many actual networks have community structures which is intense inside and relaxed outside.It is important to understand the network structure and analyze the network characteristics that understanding the structure of the community in complex networks.Community structure analysis has a very wide application in sociology,computer graphics,biology and physics,and community detection algorithm is an important means to find the structure of these societies,so the research on community detection algorithm is very meaningful and necessary.Although the time complexity of extreme value optimization based on modularity is low and fast,it can not guarantee that the global optimal solution can obtained.This kind of algorithm can not find the smaller and even important community structure in complex network.This paper aims to solve the problem that the community structure is too large due to the resolution limit problem and can not find the weakness of the small community structure in the large-scale network.And using the algorithm framework of hidden community structure in the real network Hicode(Hidden Community Detection),a community detection algorithm for mining multi-layer recursive leaf structure is proposed.In this paper,we extract the sub-networks corresponding to each explicit association and implicit association in the original sub-network,and use the Mod algorithm to further extract the community structure in the sub-network,so as to get the multi-layer sub-community structure.Through the comparison experiment,the proposed recursive community structure mining algorithm solves the defects of the modular optimization algorithm,effectively finds the weak community structure hidden under the strong community,and also is beyond the existing classic community detection algorithm in many data sets.