gptk (version 1.06)

modelGradient: Model log-likelihood/objective error function and its gradient.

Description

modeGradient gives the gradient of the objective function for a model. By default the objective function (modelObjective) is a negative log likelihood (modelLogLikelihood).

Usage

modelObjective(params, model, ...)
modelLogLikelihood(model)
modelGradient(params, model, ...)

Arguments

params

parameter vector to evaluate at.

model

model structure.

...

optional additional arguments.

Value

• g
  the gradient of the error function to be minimised.
• v
  the objective function value (lower is better).
• ll
  the log-likelihood value.

See Also

modelOptimise.