A0910
Title: Joint spectral clustering in multilayer degree-corrected stochastic blockmodels
Authors: Joshua Agterberg - Johns Hopkins University (United States)
Zachary Lubberts - University of Virginia (United States)
Jesus Arroyo - Texas A&M University (United States) [presenting]
Abstract: Modern network datasets often have multiple layers, either as different views, time-varying observations, or independent sample units. These data require models and methods that are flexible enough to capture local and global differences across the networks while simultaneously being parsimonious and tractable to yield computationally efficient and theoretically sound solutions capable of aggregating information across the networks. The multilayer degree-corrected stochastic blockmodel is considered where a collection of networks share the same community structure, but degree-corrections and block connection probability matrices are permitted to be different. The identifiability of this model is established, and a spectral clustering algorithm is proposed for community detection in this setting. The theoretical results demonstrate that the misclustering error rate of the algorithm improves exponentially with multiple network realizations, even in the presence of significant layer heterogeneity. Simulation studies show this approach improves existing multilayer community detection methods in this challenging regime. Furthermore, in a case study of US airport data from January 2016 - September 2021, it is found that this methodology identifies meaningful community structure and trends in airport popularity influenced by pandemic impacts on travel
