Usage
fitsvar.sb.iso(esv, dk = ncol(esv$data$x), nx = NULL,
rmax = esv$grid$max, min.contrib = 10,
method = c("cressie", "equal", "npairs", "linear"),
iter = 10, tol = sqrt(.Machine$double.eps))Arguments
esv
pilot semivariogram estimate, a
np.svar-class (or
svar.bin) object. Typically an output of
the funct dk
dimension of the kappa function (dk == 0
corresponds to a model valid in any dimension; if
dk > 0, it should be greater than or equal to the
dimension of the spatial process
ncol(esv$data$x)).
nx
number of discretization nodes. Defaults to
min(nesv - 1, 50), where nesv is the number
of semivariogram estimates.
rmax
maximum lag considered in the discretization
(range of the fitted variogram on output).
min.contrib
minimum number of (equivalent)
contributing pairs (pilot estimates with a lower number
are ignored, with a warning).
method
string indicating the WLS fitting method to
be used (e.g. method = "cressie"). See "Details"
below.
iter
maximum number of interations of the WLS
algorithm (used only if method == "cressie").
tol
absolute convergence tolerance (used only if
method == "cressie").