FRK: An R Package for Spatial and Spatio-Temporal Prediction with Large\n Datasets

2017/05/23 by Andrew Zammit‐Mangion, Noel Cressie, Zammit-Mangion, Andrew +1 · 7 citations
Economics, Econometrics and Finance · Environmental Science · #Computation (stat.CO) #FOS: Computer and information sciences #Land Use and Ecosystem Services #Soil Geostatistics and Mapping #Spatial and Panel Data Analysis

paper · pdf · doi:10.48550/arxiv.1705.08105

openalex publication_date 2017/05/23 · openalex created_date 2022/08/24 · openalex updated_date 2026/07/28

Abstract

FRK is an R software package for spatial/spatio-temporal modelling and\nprediction with large datasets. It facilitates optimal spatial prediction\n(kriging) on the most commonly used manifolds (in Euclidean space and on the\nsurface of the sphere), for both spatial and spatio-temporal fields. It differs\nfrom many of the packages for spatial modelling and prediction by avoiding\nstationary and isotropic covariance and variogram models, instead constructing\na spatial random effects (SRE) model on a fine-resolution discretised spatial\ndomain. The discrete element is known as a basic areal unit (BAU), whose\nintroduction in the software leads to several practical advantages. The\nsoftware can be used to (i) integrate multiple observations with different\nsupports with relative ease; (ii) obtain exact predictions at millions of\nprediction locations (without conditional simulation); and (iii) distinguish\nbetween measurement error and fine-scale variation at the resolution of the\nBAU, thereby allowing for reliable uncertainty quantification. The temporal\ncomponent is included by adding another dimension. A key component of the SRE\nmodel is the specification of spatial or spatio-temporal basis functions; in\nthe package, they can be generated automatically or by the user. The package\nalso offers automatic BAU construction, an expectation-maximisation (EM)\nalgorithm for parameter estimation, and functionality for prediction over any\nuser-specified polygons or BAUs. Use of the package is illustrated on several\nspatial and spatio-temporal datasets, and its predictions and the model it\nimplements are extensively compared to others commonly used for spatial\nprediction and modelling.\n

Citations

• Multivariable geostatistics in S: the gstat package

Cited by

• deepspat: An R package for modeling nonstationary spatial and spatio-temporal Gaussian and extremes data through deep deformations
• Three Distributional Approaches for PM10 Assessment in Northern Italy
• Multivariate Low-Rank State-Space Model with SPDE Approach for High-Dimensional Data
• A comparison between geostatistical and machine learning models for spatio-temporal prediction of PM2.5 data
• Bayesian Inference for Big Spatial Data Using Non-stationary Spectral Simulation
• Deep Gaussian Markov Random Fields
• Two-stage MCMC for Fast Bayesian Inference of Large Spatio-temporal Ordinal Data, with Application to US Drought

Related

• Modelling Big, Heterogeneous, Non-Gaussian Spatial and Spatio-Temporal Data using FRK
• The Spatial Kernel Predictor based on Huge Observation Sets
• Spatial modelling with R-INLA: A review
• A Comparison of Spatial Predictors when Datasets Could be Very Large
• Fast Spatial Prediction from Inhomogeneously Sampled Data Based on Generalized Random Fields with Gibbs Energy Functionals