This function generates another function to be used within optim to
obtain maximum likelihood estimates of covariance (and possibly mean) parameters.
The function includes options for
(1) maximum likelihood ("ml") vs. restricted maximum likelihood
("reml"),
(2) smoothness (kappa): models without smoothness vs. estimating the
smoothness vs. using fixed smoothness,
(3) locally isotropic vs. locally anisotropic, and
(4) fixed nugget variance (tausq): fixed vs. estimated.
make_local_lik(
locations,
cov.model,
data,
Xmat,
nugg2.var = matrix(0, nrow(locations), nrow(locations)),
tausq = 0,
kappa = 0.5,
fixed = rep(FALSE, 6),
method = "reml",
local.aniso = TRUE,
fix.tausq = FALSE,
fix.kappa = FALSE
)A matrix of locations.
String; the covariance model.
A vector or matrix of data to use in the likelihood calculation.
The design matrix for the mean model.
Fixed values for the variance/covariance of the second nugget term; defaults to a matrix of zeros.
Scalar; fixed value for the nugget variance (when
fix.tausq = TRUE).
Scalar; fixed value for the smoothness (when fix.kappa = TRUE).
Logical vector of FALSE values; length corresponds to the number
of parameters to be estimated.
Indicates the estimation method, either maximum likelihood ("ml")
or restricted maximum likelihood ("reml").
Logical; indicates if the local covariance should be
anisotropic (TRUE) or isotropic (FALSE). Defaults to TRUE.
Logical; indicates whether the default nugget term
(tau^2) should be fixed (TRUE) or estimated (FALSE). Defaults to
FALSE.
Logical; indicates if the kappa parameter should be
fixed (TRUE) or estimated (FALSE). Defaults to FALSE
(only valid for cov.model = "matern" and cov.model = "cauchy").
This function returns another function for use in optim.
# NOT RUN {
make_local_lik( locations, cov.model, data, Xmat )
# }
# NOT RUN {
# }
Run the code above in your browser using DataLab