ngp.m

% An NGP is a nonstationary Gaussian process model that models the
% nonstationary observed real dynamics by employing a local
% parameterizarion across an input space.
%
% NGP term is defined as part of publication of an
% algorithm LISAL that learns NGPs efficiently. NGP is also called iNGP-
% intuitive Nonstationary Gaussian Process. See documentation on lisal for
% more details.
%
%Here is a list of models that are considered part of iNGP class and
%implemented as part of this code release.
%
%HGP- Heteroscedastic Gaussian Process
%[Gibbs, 1997; Goldberg et al., 1998; MacKay, 1998]
%
%GPPM- Gaussian Process Product Model
%[MacKay, 1998; Adams & Stegle, 2008]
%
%PCLSK- Process Convolution with Local Smoothing Kernels
%[Gibbs, 1997; MacKay, 1998; Higdon, 1998; Paciorek & Schervish, 2004;
%Garg, Singh, & Ramos, 2012b]
%
%BWGP- Bayesian Warped Gaussian Process
%[Snelson, E., Rasmussen, C. E., & Ghahramani, Z. (2004)]
%
%SDIS- Spatial Deformation of Input Space
%[Schmidt, A., & O’Hagan, A. (2003)]
%
%LEIS- Latent Extension of Input space
%[Pfingsten, T., Kuss, M., & Rasmussen, C. E. (2006)]
% 
% Also see documentation on lisal.