An Eigenmodel for Dynamic Multilayer Networks
Authors: Joshua Daniel Loyal, Yuguo Chen; Published in 2023, Volume 24(128), Pages 1-69
Abstract
Dynamic multilayer networks often represent the structure of multiple co-evolving relations. However, statistical models for this type of network are not well-developed. In this study, we propose a new latent space model for dynamic multilayer networks. Our model has the unique ability to identify common time-varying structures shared by all layers, while also accounting for layer-wise variation and degree heterogeneity. We establish the identifiability of the model’s parameters and develop a structured mean-field variational inference approach to estimate the model’s posterior. This estimation procedure scales to networks that were previously considered intractable for dynamic latent space models. To validate our model, we assess its accuracy and scalability on simulated networks. Additionally, we apply the model to two real-world problems: discerning regional conflicts in a dataset of international relations and quantifying the spread of infectious diseases in a school based on students’ daily contact patterns.
[abs]