Evaluate the negative log-likelihood of a NSGEV model

Evaluate the negative log-likelihood of a NSGEV model.

negLogLik(psi, model, data = NULL, y = NULL, deriv = TRUE, checkNames = TRUE)

Arguments

psi: Vector of NSGEV parameters.
model: The NSGEV model.
data: A data frame with the covariates to be used. By default, the data attached to model is udes.
y: The response as in NSGEV. By default the response attached to the model is used.
deriv: Logical. When TRUE, the gradient \(-\partial \ell/\partial \boldsymbol{\psi}\) of the negative log-likelihood w.r.t. the parameters of the model will be returned. In this case, a list will be returned.
checkNames: Logical. If TRUE, the names of the vector psi are checked against the parnames of the model.

The negative log-likelihood (to be minimized) or a list with two elements corresponding to the value of the log-likelihood and its gradient.

The returned value of this function when deriv is TRUE is designed for use with the nloptr package.

The class of the result may be changed in future versions.

Yves Deville