A Class of Bayesian Community Detection Algorithms Based on Covariates
In the study of complex network systems,community detection is one of the important methods to mine the internal structure information of network data. This paper proposes a Bayesian community detection algorithm for a Degree-corrected Stochastic Block Model containing node covariate information ,by selecting important node covariates based on the Random Forest algorithm ,applying a covariate-dependent random partition model divide the community membership ,apply the DCSBM model ,and use the Gibbs sampler method to detect the community structure and infer the number of commu-nities. Numerical simulation results show that the Bayesian community detection algorithm for a Degree-corrected Stochastic Block Model containing node covariate information can significantly improve the performance of community detection. Finally, the Bayesian community detection algorithm for a Degree-corrected Stochastic Block Model containing node covariate infor-mation method is applied to the real data of the Weddell Sea trophic network for empirical analysis.
万方数据
维普
