CMStatistics 2021: Start Registration
View Submission - CMStatistics
Title: Modeling massive multivariate spatial data with the basis graphical lasso Authors:  Mitchell Krock - Rutgers University (United States)
William Kleiber - University of Colorado (United States) [presenting]
Dorit Hammerling - Colorado School of Mines (United States)
Stephen Becker - University of Colorado Boulder (United States)
Abstract: A new modeling framework is proposed for highly multivariate spatial processes that synthesizes ideas from recent multiscale and spectral approaches with graphical models. The basis graphical lasso writes a univariate Gaussian process as a linear combination of basis functions weighted with entries of a Gaussian graphical vector whose graph is estimated from optimizing an $L_1$ penalized likelihood. We extend the setting to a multivariate Gaussian process where the basis functions are weighted with Gaussian graphical vectors. We motivate a model where the basis functions represent different levels of resolution, and the graphical vectors for each level are assumed to be independent. Using an orthogonal basis grants linear complexity and memory usage in the number of spatial locations, the number of basis functions, and the number of realizations. An additional fusion penalty encourages a parsimonious conditional independence structure in the multilevel graphical model. We illustrate our method on a large climate ensemble from the National Center for Atmospheric Research's Community Atmosphere Model that involves 40 spatial processes.