Optimal rates for community estimation in the weighted stochastic block model (Q2176614)

An important property of stochastic block methods (SBM) is that all edges are assumed to be binary. In contrast, the edges appearing in many real-world networks are weighted. It justified to study weighted SBM. Each edge is generated from a Bernoulli(\(p\)) or Bernoulli(\(q\)) distribution depending on whether its endpoints lie in the same community. The main theoretical contribution is to characterize the optimal rate of misclustering error in the weighted SBM. The results show that the optimal rate for community estimator in a weighted SBM is governed by the Renyi-divergence of order \(\frac{1}{2}\) between two mixed distributions capturing the discrepancy between the edge probabilities and edge weight densities for between-community and within-community connections.