Community Detection In Signed Networks And Dynamic Networks Based On Discrete-time Model

Community detection of complex networks is a widely concerned direction in the research of complexity science,and has significant contribution and sustained impact in the fields of information science,biology,mathematics and sociology.In recent years,for different types of complex networks,people have proposed a lot of algorithms to find community structure,known as community detection algorithm.Based on the universality of community detection in today’s society,community detection in signed network based on similarity and community detection in dynamic network are taken as the main research content in this paper.This paper studies how to define a reasonable similarity model according to the characteristics of signed networks and how to quantify the real-time information of the dynamic network to establish a reasonable mathematical model,expressed as follows:(1)Community detection in signed networks based on discrete-time model.In this paper,consider the characteristics of positive and negative connections in signed networks,we defined a new similarity formula.Added the similarity to the dynamic evolution model,so that the nodes’ state in the signed network evolves according to the network model.The theory proved that the model can reach Lyapunov stability.Based on the simulations of real networks and synthetic networks,and compared with the existing algorithm,the algorithm results are superior to the existing algorithms in terms of time and accuracy.(2)Community detection in dynamic networks based on discrete-time model.In this paper,for the characteristics of dynamic network changes over time,consider the network structure of previous time step and the network structure of current time step,we get a new adjacent matrix by weighting the network adjacent matrix of different time steps.The time-dependent adjacent matrix and the dynamic network model is applied to realize the community detection in dynamic networks.The experimental results shown that the algorithm is applicable not only to small-scale dynamic networks,but also to large-scale dynamic networks with large number of nodes and unbalancedcommunity structure.