Consistency of Spectral Clustering on Hierarchical Stochastic Block Models

Abstract

We study the hierarchy of communities in real-world networks under a generic stochastic block model, in which the connection probabilities are structured in a binary tree. Under such model, a standard recursive bi-partitioning algorithm is dividing the network into two communities based on the Fiedler vector of the unnormalized graph Laplacian and repeating the split until a stopping rule indicates no further community structures. We prove the strong consistency of this method under a wide range of model parameters, which include sparse networks with node degrees as small as $O(\log n)$. In addition, unlike most of existing work, our theory covers multiscale networks where the connection probabilities may differ by orders of magnitude, which comprise an important class of models that are practically relevant but technically challenging to deal with. Finally we demonstrate the performance of our algorithm on synthetic data and real-world examples.

45 pages, 7 figures

References in corpus (3)

• Hierarchical structure and the prediction of missing links in networks
• Unperturbed: spectral analysis beyond Davis-Kahan
• Unified ${\ell }_{2\to \infty }$ Eigenspace Perturbation Theory for Symmetric Random Matrices

Cited by in corpus (6)

• Linear regression and its inference on noisy network-linked data
• Informative core identification in complex networks
• The Importance of Being Correlated: Implications of Dependence in Joint Spectral Inference across Multiple Networks
• Community models for networks observed through edge nominations
• Distributed Community Detection for Large Scale Networks Using Stochastic Block Model
• Graph Embedding with Hierarchical Attentive Membership