Extracting Community Structures in Complex Networks Based on Discrete Neural Network Algorithm

Abstract

The community structure in a complex network is a subset of nodes whose internal nodes are closely connected and relatively sparsely connected to other parts of the network. The community structure is a key structural law of complex networks, so the association of complex networks is accurately analyzed. Structure is a very important topic in the study of complex networks.This paper studies the problem of complex network community structure extraction, and proposes the DHNN algorithm. This algorithm shows that from the arbitrary initial value, after several iterations, it finally converges to an attractor or a limit loop of length 2, giving the energy of DHNN. The relationship between function and modularity function proves that the stable point of the network corresponds to a very large modular function Q value, and the example verification work is carried out. For the DHNN algorithm proposed in this paper, the simulation experiment is carried out on the actual network. The results show that the eigenvalue eigenvector algorithm of DHNN community structure extraction algorithm Newman has a large Q value. At the same time, the DHNN algorithm does not need to calculate the eigenvalue eigenvectors, and only needs a simple addition multiplication operation to extract the community structure in the network. Therefore, the proposed algorithm, especially the DHNN algorithm, has powerful computing power. Community structures in larger, complex networks can be extracted.