Learn R Programming

stilt (version 1.3.0)

Separable Gaussian Process Interpolation (Emulation)

Description

Functions for simplified emulation of time series computer model output in model parameter space using Gaussian processes. Stilt can be used more generally for Kriging of spatio-temporal fields. There are functions to predict at new parameter settings, to test the emulator using cross-validation (which includes information on 95% confidence interval empirical coverage), and to produce contour plots over 2D slices in model parameter space.

Copy Link

Version

Install

install.packages('stilt')

Monthly Downloads

499

Version

1.3.0

License

GPL-3

Maintainer

Kelsey Ruckert

Last Published

August 15th, 2018

Functions in stilt (1.3.0)

stilt-internal

Internal Stilt Functions
Data.AR1Korea.model

Korean modelled temperature variability and future change
Data.AR1Korea.par

Korean temperature variability parameters
Data.Sicopolis.model

SICOPOLIS model ensemble output
Data.UVic.model

UVic ESCM climate model ensemble output
predict.emul

Predict Method for an Emulator
Data.Sicopolis.par

SICOPOLIS model ensemble parameter settings
Data.UVic.par

UVic ESCM climate model ensemble parameter settings
rsurface.plot

To produce a response surface plot of the emulator
emul.predict

To predict using an emulator (Deprecated)
test.csv

To cross-validate a Gaussian Process emulator
stilt-package

Separable Gaussian Process Emulator
test.all

To test an emulator using leave-one-out cross-validation
emulator

Function to fit an emulator to ensemble model output
Data.1D.model

Synthetic model output for a simple 1-parameter ensemble example
Data.1D.par

Parameter settings for 1-parameter ensemble model output in Data.1D.model
emul.Sicopolis

SICOPOLIS ice sheet model emulator
sep.cov

Construct time and parameter covariance matrices
emul.1D

Emulator of model for the 1-parameter example