A1476
Title: Nonparametric estimation for time-varying network models
Authors: Tianxi Li - University of Minnesota (United States) [presenting]
Adam Rothman - University of Minnesota (United States)
Jeonghwan Lee - University of Minnesota, Twin Cities (United States)
Abstract: The analysis of time-varying networks, in which interactions evolve over time, poses significant statistical challenges. We address the problem of estimating time-varying edge probabilities within the flexible framework of the time-varying graphon model, which posits proper smooth evolutions in both latent node positions and time. We introduce a stagewise smoothing method to estimate a model that is computationally efficient with a uniform convergence guarantee. The method is demonstrated by the modeling evaluation of the US Congress network.
