deepgp Package

Performs model fitting and sequential design for deep Gaussian processes using MCMC and elliptical slice sampling. Models extend up to three layers deep; a one layer model is equivalent to typical Gaussian process regression. Sequential design criteria include integrated mean square prediction error (IMSPE), active learning Cohn (ALC), and expected improvement (EI). Covariance structure is based on inverse exponentiated squared euclidean distance. Applicable to noisy and deterministic functions.
Incorporates SNOW parallelization and utilizes C under the hood. Manuscript forthcoming.

View deepgp-package help file for more information.

CRAN submission notes:

This is a re-submission. The following changes were made: * Added references to description field of DESCRIPTION file * Added toy examples that run in less than 5 seconds to ?fit_one_layer, ?fit_two_layer, and ?fit_three_layer documentation * Implemented on.exit to restore par options after plotting

Test environments

R CMD check results

There were no ERRORs or WARNINGs or NOTEs.