Penalized basis models for very large spatial datasets

TL;DR

This paper introduces a flexible, scalable spatial modeling approach using penalized graphical models for basis coefficients, enabling analysis of very large datasets with improved model fit and interpretability.

Contribution

It proposes a novel penalized likelihood framework for modeling stochastic coefficients in basis expansions, connecting to graphical lasso for efficient computation.

Findings

Successfully applied to large climatology datasets

Achieves better AIC scores than LatticeKrig

Provides interpretable graphical structures

Abstract

Many modern spatial models express the stochastic variation component as a basis expansion with random coefficients. Low rank models, approximate spectral decompositions, multiresolution representations, stochastic partial differential equations and empirical orthogonal functions all fall within this basic framework. Given a particular basis, stochastic dependence relies on flexible modeling of the coefficients. Under a Gaussianity assumption, we propose a graphical model family for the stochastic coefficients by parameterizing the precision matrix. Sparsity in the precision matrix is encouraged using a penalized likelihood framework. Computations follow from a majorization-minimization approach, a byproduct of which is a connection to the graphical lasso. The result is a flexible nonstationary spatial model that is adaptable to very large datasets. We apply the model to two large and heterogeneous spatial datasets in statistical climatology and recover physically sensible graphical structures. Moreover, the model performs competitively against the popular LatticeKrig model in predictive cross-validation, but substantially improves the Akaike information criterion score.