Recovery, detection and confidence sets of communities in a sparse stochastic block model
share
Posterior distributions for community assignment in the planted bi-section model are shown to achieve frequentist exact recovery and detection under sharp lower bounds on sparsity. Assuming posterior recovery (or detection), one may interpret credible sets (or enlarged credible sets) as consistent confidence sets. If credible levels grow to one quickly enough, credible sets can be interpreted as frequentist confidence sets without conditions on the parameters. In the regime where within-class and between-class edge-probabilities are very close, credible sets may be enlarged to achieve frequentist asymptotic coverage. The diameters of credible sets are controlled and match rates of posterior convergence.
READ FULL TEXT
