https://github.com/cran/RandomFields
Tip revision: fd4911aa390fd49ddab92bd139bbbf35422e32e5 authored by Martin Schlather on 06 February 2020, 05:20:37 UTC
version 3.3.8
version 3.3.8
Tip revision: fd4911a
RPgauss.Rd
\name{RPgauss}
\alias{RPgauss}
\title{Simulation of Gaussian Random Fields}
\description{
This function is used to specify a Gaussian random field that
is to be simulated or estimated.
Returns an object of class \code{\link[=RMmodel-class]{RMmodel}}.
}
\usage{
RPgauss(phi, boxcox, stationary_only)
}
\arguments{
\item{phi}{the \command{\link{RMmodel}}.}
% \item{loggauss}{logical. If \code{TRUE} then a log-Gaussian random
% field is returned. Default is \code{FALSE}.
% }
\item{boxcox}{the one or two parameters of the box cox transformation.
If not given, the globally defined parameters are used.
See \command{\link{RFboxcox}} for details.
}
\item{stationary_only}{Logical or NA. Used for the automatic
choice of methods.
\itemize{
\item \code{TRUE}: The simulation of non-stationary random
fields is refused. In particular, the intrinsic
embedding method is excluded and
the simulation of Brownian motion is rejected.
\item \code{FALSE}: Intrinsic embedding is always allowed;
actually, it's the first one considered in the automatic
selection algorithm.
\item \code{NA}: The simulation of the Brownian motion is allowed,
but intrinsic embedding is not used for translation invariant
(\dQuote{stationary}) covariance models.
}
Default: \code{NA}.
}
}
\value{
The function returns an object of class \code{\link[=RMmodel-class]{RMmodel}}.
}
\note{
In most cases, \code{RPgauss} need not be given explicitly as
Gaussian random fields are assumed as default.
\command{\link{RPgauss}} may not find the fastest
method neither the most precise one. It just finds any method
among the available methods. (However, it guesses what is a good
choice.)
See \command{\link{RFgetMethodNames}} for further information.
Note that some of the methods do not work for all covariance
or variogram models, see \code{\link{RFgetModelNames}(intern=FALSE)}.
By default, all Gaussian random fields have zero mean.
Simulating with trend can be done by including \command{\link{RMtrend}}
in the model.
\command{\link{RPgauss}} allows to simulate different classes of random fields,
controlled by the wrapping model:
If the submodel is a pure covariance or
variogram model, i.e. of class \code{\link[=RMmodel-class]{RMmodel}}, a
corresponding centered Gaussian field is simulated.
Not only stationary fields but also non-stationary and anisotropic
models can be used, e.g. zonal anisotropy, geometrical
anisotropy, separable models, non-separable space-time models,
multiplicative or nested models;
see \command{\link{RMmodel}} for a list of all available models.
}
\me
\seealso{
\link{RP},
\link{Gaussian},
\command{\link{RMmodel}},
\command{\link{RFoptions}},
\command{\link{RPbrownresnick}},
\command{\link{RPchi2}},
\command{\link{RPopitz}},
\command{\link{RPt}},
\command{\link{RPschlather}}.
Do not mix up with \command{\link{RMgauss}} or \command{\link{RRgauss}}.
}
\keyword{spatial}
\examples{\dontshow{StartExample()}
RFoptions(seed=0) ## *ANY* simulation will have the random seed 0; set
## RFoptions(seed=NA) to make them all random again
model <- RMexp()
x <- seq(0, 10, 0.02)
plot(model)
plot(RFsimulate(model, x=x, seed=0))
plot(RFsimulate(RPgauss(model), x=x, seed=0), col=2) ## the same
\dontshow{FinalizeExample()}}