lineqGPR (version 0.1.1)

logLikAdditiveGrad: Gradient of the Log-Likelihood of a Additive Gaussian Process.

Description

Compute the gradient of the negative log-likelihood of an Additive Gaussian Process.

Usage

logLikAdditiveGrad(
  par = unlist(purrr::map(model$kernParam, "par")),
  model,
  parfixed = rep(FALSE, model$d * length(par)),
  mcmc.opts = NULL,
  estim.varnoise = FALSE
)

Arguments

par

the values of the covariance parameters.

model

an object with "lineqAGP" S3 class.

parfixed

indices of fixed parameters to do not be optimised.

mcmc.opts

not used.

estim.varnoise

If true, a noise variance is estimated.

Value

the gradient of the negative log-likelihood.

References

Rasmussen, C. E. and Williams, C. K. I. (2005), "Gaussian Processes for Machine Learning (Adaptive Computation and Machine Learning)". The MIT Press. [link]

See Also

logLikAdditiveFun