gpmlr

gpmlr allows the GPML (Gaussian Processes for Machine Learning) toolbox Matlab function for Gaussian process inference and prediction to be called from R, using R syntax and data structures.

Installation

gpmlr is not on CRAN. You can install the development version from GitHub via

remotes::install_github("duckmayr/gpmlr")

Installation and use on Windows systems is not yet supported.

Please also note that before installing gplmr, you must install Octave. You can obtain Octave from their website, https://www.gnu.org/software/octave/.

If you are on a Debian based system such as Ubuntu, you can also install Octave using your package manager

sudo apt update
sudo apt install octave liboctave-dev

On Mac, you can use Homebrew to install Octave:

brew update
brew install octave

We do provide the GPML code so that anyone with Octave installed can use the R package out of the box.

Using gpmlr

gpmlr provides an R facing function, gp(), that calls GPML's gp() MATLAB function. Some basic documentation for our R interface is available via

help("gp")

You can also see an example of its use via

example(gp)

However, we have not yet added extensive documentation on all the inference methods or mean, covariance, and likelihood functions available for use in gp() from GPML. Documentation on these functions will be added in a vignette, but in the meantime is available in the Matlab files themselves under inst/gpml.

Contributing

Before contributing, please consult the contributing guidelines in CONTRIBUTING.md.

License

GPL (>= 2)