gptk (version 1.0)

gpGradient: Gradient wrapper for a GP model.

Description

wraps the log likelihood gradient function to return the gradient of the negative of the log likelihood. This can then be used in, for example, NETLAB, minimisation tools.

Usage

g <- gpGradient(params, model)

Arguments

params

the parameters of the model.

model

the model for which gradients will be computed.

Value

• g
  the returned gradient of the negative log likelihood for the given parameters.

See Also

gpCreate, gpGradient, gpLogLikeGradients, gpOptimise.

Examples

## missing