Community detection with dependent connectivity

In network analysis, within-community members are more likely to be connected than between-community members, which is reflected in that the edges within a community intercorrelated. However, existing probabilistic models for detection such as stochastic block model (SBM) not designed capture dependence among edges. this paper, we propose new approach incorporate intra-community of connectivities through Bahadur representation. The proposed method does require specifying likelihood function, could intractable correlated binary connectivities. addition, allows heterogeneity between different communities. theory, show incorporating correlation information can achieve faster convergence rate compared independent SBM, and algorithm has lower estimation bias accelerated variational EM. Our simulation studies outperforms multinetwork methods assuming conditional independence We also demonstrate application agricultural product trading networks from countries brain fMRI imaging networks.