Efficient Emulators of Computer Experiments Using Compactly Supported Correlation Functions

This is the R code to carry out the examples in the paper of the above title by myself, Derek Bingam, et al. The main code is stored in an (as yet) undocumented R package, to aid in loading external Fortran code. To see the main functions, install the package and try the code in SparseEmExample.R. If you want to go “under the hood,” have a look at the file SparseEmFunctions.R, which contains all my comments for functions in the package.

SparseEm_0.2-2.tar.gz
SparseEmExample.R
SparseEmFunctions.R

Bayesian Functional ANOVA Modelling Using Gaussian Process Priors

This is the R code to carry out the examples in the paper of the above title by myself and Stephan Sain. Each zipped file contains both data and code. Please read the README files for instructions.

Canadian temperature example
Regional climate models example

Estimating Spatial Covariance Parameters in Precipitation Anomalies Using Covariance Tapering

R code and data for the example in Kaufman, Schervish, and Nychka (2008) is archived here.

R Package ‘sphere’

The sphere package is object-oriented code for working with objects consisting of observations on the sphere, ie., unit vectors in 3D. There are functions for manipulating these objects, plotting projections, and some basic statistical analyses. Here's a representative plot, showing some spherical data (point locations represent orientation and color the associated data value) and a smoothed surface.

spherekernel

Download package
Example file

Last updated August, 2011.