gprLikelihood: Calculate negative log-likelihood

In martinzaefferer/COBBS: Continuous Optimization Benchmarks By Simulation

Calculate negative log-likelihood

Description

Used to determine theta/lambda values for the GPR model in gaussianProcessRegression. Supportive function for GPR model, not to be used directly.

Usage

gprLikelihood(x, AX, Ay, optimizeP = FALSE, useLambda = TRUE, penval = 1e+08)

Arguments

x	vector, containing parameters log10(theta), log10(lambda) and p.
AX	3 dimensional array, constructed by gaussianProcessRegression from the sample locations
Ay	vector of observations at sample locations
optimizeP	boolean, whether to optimize parameter p (exponents) or fix at two.
useLambda	boolean, whether to use nugget
penval	a penalty value which affects the value returned for invalid correlation matrices / configurations

Value

list with elements
NegLnLike concentrated log-likelihood *-1 for minimising
Psi correlation matrix
Psinv inverse of correlation matrix (to save computation time during prediction)
mu
ssq

See Also

gaussianProcessRegression

martinzaefferer/COBBS documentation built on July 19, 2023, 4:12 a.m.