Learn R Programming

⚠️There's a newer version (0.5.1) of this package.Take me there.

GpGp

GpGp is an R package for fast approximate Gaussian process computation. The package includes implementations of the Vecchia's (1988) original approximation, as well as several updates to it, including the reordered and grouped versions of the approximation outlined in Guinness (2018).

Installing

The package can be installed from CRAN with the usual R command

install.packages("GpGp")

or directly from Github for the latest version

devtools::install_github("joeguinness/GpGp")

We always recommend using multithreaded linear algebra libraries in R, but for this package in particular, using multithreaded libraries can have a big impact on performance. On a Mac, there is a very simple way to link to the Apple Accelerate Framework. On PC and Linux, it's more complicated, but you can use Microsoft R Open instead, which comes automatically with multithreaded libraries.

Basic Use

The main function for fitting models is called 'fit_model', and the main function for doing predictions is called 'predictions'.

See this youtube video for a tutorial: https://www.youtube.com/watch?v=phyB4n0CDWg&t=4s

Copy Link

Version

Install

install.packages('GpGp')

Monthly Downloads

475

Version

0.3.0

License

MIT + file LICENSE

Maintainer

Joseph Guinness

Last Published

October 3rd, 2020

Functions in GpGp (0.3.0)

Multiply transpose of approximate Cholesky by a vector

Ocean temperatures from Argo profiling floats

Multiply approximate Cholesky by a vector

expit function and integral of expit function

Multiply transpose of approximate inverse Cholesky by a vector

exponential_anisotropic2D

Geometrically anisotropic exponential covariance function (two dimensions)

GpGp: Fast Gaussian Process Computing.

Multiply approximate inverse Cholesky by a vector

Conditional Simulation using Vecchia's approximation

condition_number

compute condition number of matrix

exponential_anisotropic3D

Geometrically anisotropic exponential covariance function (three dimensions)

exponential_isotropic

Isotropic exponential covariance function

exponential_sphere_warp

Deformed exponential covariance function on sphere

exponential_nonstat_var

Isotropic exponential covariance function, nonstationary variances

exponential_scaledim

Exponential covariance function, different range parameter for each dimension

Fisher scoring algorithm

find_ordered_nn_brute

Naive brute force nearest neighbor finder

fast_Gp_sim_Linv

Approximate GP simulation with specified Linverse

exponential_spheretime

Exponential covariance function on sphere x time

exponential_spacetime

Spatial-Temporal exponential covariance function

exponential_sphere

Isotropic exponential covariance function on sphere

find_ordered_nn

Find ordered nearest neighbors.

Windspeed measurements from Jason-3 Satellite

matern15_isotropic

Isotropic Matern covariance function, smoothness = 1.5

Automatic grouping (partitioning) of locations

exponential_spheretime_warp

Deformed exponential covariance function on sphere

matern_anisotropic2D

Geometrically anisotropic Matern covariance function (two dimensions)

Estimate mean and covariance parameters

Approximate GP simulation

matern_nonstat_var

Isotropic Matern covariance function, nonstationary variances

matern_isotropic

Isotropic Matern covariance function

matern45_isotropic

Isotropic Matern covariance function, smoothness = 4.5

matern_anisotropic3D

Geometrically anisotropic Matern covariance function (three dimensions)

matern45_scaledim

Matern covariance function, smoothess = 3.5, different range parameter for each dimension

get link function, whether locations are lonlat and space time

penalize small values of parameter: penalty, 1st deriative, 2nd derivative

matern_sphere_warp

Deformed Matern covariance function on sphere

Isotropic Matern covariance function on sphere

penalize large values of parameter: penalty, 1st deriative, 2nd derivative

matern15_scaledim

Matern covariance function, smoothess = 1.5, different range parameter for each dimension

order_dist_to_point

Distance to specified point ordering

order_coordinate

Sorted coordinate ordering

matern_scaledim

Matern covariance function, different range parameter for each dimension

matern35_isotropic

Isotropic Matern covariance function, smoothness = 3.5

matern35_scaledim

Matern covariance function, smoothess = 3.5, different range parameter for each dimension

matern_spacetime

Spatial-Temporal Matern covariance function

order_middleout

Middle-out ordering

Maximum minimum distance ordering

get penalty function

vecchia_meanzero_loglik

Vecchia's approximation to the Gaussian loglikelihood, zero mean

vecchia_grouped_profbeta_loglik_grad_info

Grouped Vecchia loglikelihood, gradient, Fisher information

matern_spheretime_warp

Deformed Matern covariance function on sphere

matern_spheretime

Matern covariance function on sphere x time

vecchia_profbeta_loglik

Vecchia's approximation to the Gaussian loglikelihood, with profiled regression coefficients.

test_likelihood_object

test likelihood object for NA or Inf values

Entries of inverse Cholesky approximation

vecchia_grouped_meanzero_loglik

Grouped Vecchia approximation to the Gaussian loglikelihood, zero mean

get_start_parms

get default starting values of covariance parameters

matern25_isotropic

Isotropic Matern covariance function, smoothness = 2.5

Compute Gaussian process predictions using Vecchia's approximations

penalize small values of log parameter: penalty, 1st deriative, 2nd derivative

vecchia_grouped_profbeta_loglik

Grouped Vecchia approximation, profiled regression coefficients

matern25_scaledim

Matern covariance function, smoothess = 2.5, different range parameter for each dimension

vecchia_profbeta_loglik_grad_info

Vecchia's loglikelihood, gradient, and Fisher information

summary.GpGp_fit

Print summary of GpGp fit

compute gradient of spherical harmonics functions