https://github.com/cran/RandomFields
Tip revision: e994a4415e67fa60cbfd3f208aaab20872521c0b authored by Martin Schlather on 14 February 2019, 21:02:19 UTC
version 3.3
version 3.3
Tip revision: e994a44
RMblend.Rd
\name{RMblend}
\alias{RMblend}
\title{Scale model for a few areas of different scales and/or
differentiabilities}
\description{
Let \eqn{Z=(Z_1, \ldots Z_k)} be an \eqn{k}-variate random field
and \eqn{A_1,\ldots, A_k} a partition of the space.
Then
\deqn{Y(x) = \sum_{i=1}^k Z_i * 1(x \in A_i)}
i.e. the model blends the components of \eqn{Z} to a new, univariate
model \eqn{Y}.
}
\usage{
RMblend(multi, blend, thresholds, var, scale, Aniso, proj)
}
\arguments{
\item{multi}{a multivariate covariance function}
\item{blend,thresholds}{The \code{threshold} is a vector of increasing
values. If the value of \code{blend} is below all thresholds up to
the \eqn{k}-th threshold, then the \eqn{k}-th component of the field
given by \code{multi} is taken. If necessary the components are recycled.
Default: \code{threshold = 0.5}, useful for blending a bivariate
field if \code{blend} takes only the values \eqn{0} and {1}.
}
\item{var,scale,Aniso,proj}{optional arguments; same meaning for any
\command{\link{RMmodel}}. If not passed, the above
covariance function remains unmodified.}
}
\value{
\command{\link{RMblend}} returns an object of class \code{\link[=RMmodel-class]{RMmodel}}.
}
\references{
\itemize{
\item Bonat, W.H. , Ribeiro, P. Jr. and Schlather, M. (2019)
Modelling non-stationarity in scale. In preparation.
\item Genton, Apanovich Biometrika.
}
}
\me
\seealso{
\command{\link{RMSadvanced}},
\command{\link{RMbubble}},
\command{\link{RMscale}},
}
\examples{\dontshow{StartExample()}
RFoptions(seed=0) ## *ANY* simulation will have the random seed 0; set
## RFoptions(seed=NA) to make them all random again
x <- seq(0,1, if (interactive()) 0.01 else 0.5)
len <- length(x)
m <- matrix(1:len, nc=len, nr=len)
m <- m > t(m)
image(m) # two areas separated by the first bisector
biwm <- RMbiwm(nudiag=c(0.3, 1), nured=1, rhored=1, cdiag=c(1, 1),
s=c(1, 1, 0.5))
model <- RMblend(multi=biwm, blend=RMcovariate(data = as.double(m), raw=TRUE))
plot(z <- RFsimulate(model, x, x)) ## takes a while ...
\dontshow{FinalizeExample()}}
\keyword{spatial}
\keyword{models}