Gaussian spatial regression using the spmoran package: case study examples

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autore principale: Murakami, Daisuke
Natura: Preprint
Pubblicazione: 2017
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866913203980599296
author Murakami, Daisuke
author_facet Murakami, Daisuke
contents This study demonstrates how to use the "spmoran" package implementing scalable spatial regression models for Gaussian and non-Gaussian data. Implemented models include spatially varying coefficient models, models with group effects, spatial unconditional quantile regression model, and low rank spatial econometric models. All of these models are estimated in a computationally efficient manner for large samples. Moran eigenvectors are used to an approximate Gaussian process (GP) modeling that is interpretable in terms of the Moran coefficient. The GP is used for modeling the spatial processes in residuals and regression coefficients. The sample codes are available from https://github.com/dmuraka/spmoran. While this vignette mainly focuses on Gaussian data modeling, another vignette focusing on non-Gaussian data including count regression is also available from the same GitHub page.
format Preprint
id arxiv_https___arxiv_org_abs_1703_04467
institution arXiv
publishDate 2017
record_format arxiv
spellingShingle Gaussian spatial regression using the spmoran package: case study examples
Murakami, Daisuke
Other Statistics
Computation
This study demonstrates how to use the "spmoran" package implementing scalable spatial regression models for Gaussian and non-Gaussian data. Implemented models include spatially varying coefficient models, models with group effects, spatial unconditional quantile regression model, and low rank spatial econometric models. All of these models are estimated in a computationally efficient manner for large samples. Moran eigenvectors are used to an approximate Gaussian process (GP) modeling that is interpretable in terms of the Moran coefficient. The GP is used for modeling the spatial processes in residuals and regression coefficients. The sample codes are available from https://github.com/dmuraka/spmoran. While this vignette mainly focuses on Gaussian data modeling, another vignette focusing on non-Gaussian data including count regression is also available from the same GitHub page.
title Gaussian spatial regression using the spmoran package: case study examples
topic Other Statistics
Computation
url https://arxiv.org/abs/1703.04467