Fast Spatial and Spatio-Temporal Regression using Moran Eigenvectors

A collection of functions for estimating spatial and spatio-temporal regression models. Moran eigenvectors are used as spatial basis functions to efficiently approximate spatially dependent Gaussian processes (i.e., random effects eigenvector spatial filtering; see Murakami and Griffith 2015 ). The implemented models include linear regression with residual spatial dependence, spatially/spatio-temporally varying coefficient models (Murakami et al., 2017, 2024; ,), spatially filtered unconditional quantile regression (Murakami and Seya, 2019 ), Gaussian and non-Gaussian spatial mixed models through compositionally-warping (Murakami et al. 2021, ).