Bayesian Learning of Graph Substructures
arXiv:2203.11664 · doi:10.1214/22-BA1338
Abstract
Graphical models provide a powerful methodology for learning the conditional independence structure in multivariate data. Inference is often focused on estimating individual edges in the latent graph. Nonetheless, there is increasing interest in inferring more complex structures, such as communities, for multiple reasons, including more effective information retrieval and better interpretability. Stochastic blockmodels offer a powerful tool to detect such structure in a network. We thus propose to exploit advances in random graph theory and embed them within the graphical models framework. A consequence of this approach is the propagation of the uncertainty in graph estimation to large-scale structure learning. We consider Bayesian nonparametric stochastic blockmodels as priors on the graph. We extend such models to consider clique-based blocks and to multiple graph settings introducing a novel prior process based on a Dependent Dirichlet process. Moreover, we devise a tailored computation strategy of Bayes factors for block structure based on the Savage-Dickey ratio to test for presence of larger structure in a graph. We demonstrate our approach in simulations as well as on real data applications in finance and transcriptomics.
arXiv comment41 pages, 8 figures
References in corpus (19)
- Natural Scales in Geographical Patterns
- Cooperative Game Theory Approaches for Network Partitioning
- Hierarchical organization of modularity in metabolic networks
- Community detection in networks: A user guide
- Optimal predictive model selection
- A Review of Stochastic Block Models and Extensions for Graph Clustering
- Nonparametric Bayesian inference of the microcanonical stochastic block model
- Sparse graphs using exchangeable random measures
- Network reconstruction and community detection from dynamics
- Clustering Network Layers With the Strata Multilayer Stochastic Block Model
- High-dimensional covariance estimation based on Gaussian graphical models
- Non-parametric Bayesian modeling of complex networks
- Bayesian inference for multiple Gaussian graphical models with application to metabolic association networks
- Inferring sparse Gaussian graphical models with latent structure
- Bayesian estimation of the latent dimension and communities in stochastic blockmodels
- The G-Wishart Weighted Proposal Algorithm: Efficient Posterior Computation for Gaussian Graphical Models
- Separate Exchangeability as Modeling Principle in Bayesian Nonparametrics
- Stochastic blockmodels for exchangeable collections of networks
- Dynamic degree-corrected blockmodels for social networks: a nonparametric approach