gpr (version 1.1)

covSEiso: Squared Exponential covariance function with isotropic distance measure.

Description

This covariance function is defined as: k(x^p,x^q) = sf2 * exp(- t(x^p - x^q)*inv(P)*(x^p - x^q)/2) , in which where the P matrix is ell^2 times the unit matrix and sf2 is the signal variance. The hyperparameter vector is loghyper = [ log(ell) ,log(sqrt(sf2)) ]

Usage

covSEiso(loghyper = NULL , x = NULL , z =  NULL , testset.covariances= FALSE)

Arguments

loghyper

loghyper is hyperparameter vector variable.

x

x is Input parameter array to apply the function over.

z

z is Index number of loghyper vector.

testset.covariances

testset.covariances is a logic value to decide to compute testset covariances or not.

Value

If z is not null and testset.covariances is TRUE this function calculates test set covariances and if its FALSE the function computes derivative matrix. When covNoise is called without parameters is reports the minimum number of parameters other than loghyper which it can accept. The output of this function is a list consisting variables A and B. B will include testset covariances calculation when testset.covariances is TRUE.

References

Carl Edward Rasmussen and Christopher K. I. Williams.Gaussian Processes for Machine Learning. MIT Press, 2006. ISBN 0-262-18253-X. Carl Edward Rasmussen & Hannes Nickisch. gpml(GAUSSIAN PROCESS REGRESSION AND CLASSIFICATION Toolbox) Matlab Library.

Examples

params= covSEiso()
params