REX-SUB: A Scalable Subsampling Strategy for Modeling Large Spatial Datasets

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Rios, Nicholas, Lee, Ben Seiyon
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910223869935616
author Rios, Nicholas
Lee, Ben Seiyon
author_facet Rios, Nicholas
Lee, Ben Seiyon
contents Recent advances in data collection technologies have led to the emergence of massive spatial datasets, with measurements obtained at millions of spatial locations. Geostatistical models typically employ Gaussian processes (GPs) to capture spatial dependence, but standard GP fitting becomes prohibitive at such scales. A promising solution is optimal subsampling, where a subset of locations is selected that optimizes a criterion. In this study, we propose a randomized exchange algorithm for subsampling (REX-SUB) which efficiently selects small subsamples that minimize prediction errors in the fitted spatial GP models. To further improve computational efficiency, we embed a scalable Vecchia approximation to the GP's joint likelihood, which takes advantage of sparsity in the precision matrix to enable fast inference on the selected subsamples. Through a simulation study and an application to a remotely sensed precipitable water dataset, we show that REX-SUB yields lower mean squared prediction errors and interval scores compared to competing subsampling strategies.
format Preprint
id arxiv_https___arxiv_org_abs_2605_16075
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle REX-SUB: A Scalable Subsampling Strategy for Modeling Large Spatial Datasets
Rios, Nicholas
Lee, Ben Seiyon
Methodology
Computation
Recent advances in data collection technologies have led to the emergence of massive spatial datasets, with measurements obtained at millions of spatial locations. Geostatistical models typically employ Gaussian processes (GPs) to capture spatial dependence, but standard GP fitting becomes prohibitive at such scales. A promising solution is optimal subsampling, where a subset of locations is selected that optimizes a criterion. In this study, we propose a randomized exchange algorithm for subsampling (REX-SUB) which efficiently selects small subsamples that minimize prediction errors in the fitted spatial GP models. To further improve computational efficiency, we embed a scalable Vecchia approximation to the GP's joint likelihood, which takes advantage of sparsity in the precision matrix to enable fast inference on the selected subsamples. Through a simulation study and an application to a remotely sensed precipitable water dataset, we show that REX-SUB yields lower mean squared prediction errors and interval scores compared to competing subsampling strategies.
title REX-SUB: A Scalable Subsampling Strategy for Modeling Large Spatial Datasets
topic Methodology
Computation
url https://arxiv.org/abs/2605.16075