Learn R Programming

RandomFields (version 3.0.32)

GSPSJ06: Fast and Exact Simulation of Large Gaussian Lattice Systems in R2

Description

Here the code of the paper on Fast and Exact Simulation of Large Gaussian Lattice Systems in R2 is given.

Arguments

References

  • Gneiting, T., Sevcikova, H., Percival, D.B., Schlather, M., Jiang, Y. (2006) Fast and Exact Simulation of Large Gaussian Lattice Systems in R2: Exploring the Limits.J. Comput. Graph. Stat.,15, 483-501.

Examples

Run this code
## Figure 1 
stabletest <- function(alpha, theta, size=if (interactive()) 512 else 20) {
  RFoptions(trials=1, tolIm = 1e-8, tolRe=0, force = FALSE,
            useprimes=TRUE, strategy=0, skipchecks=!FALSE,
             storing=TRUE)
  model <- RMcutoff(diameter=theta, a=1, RMstable(alpha=alpha))
  RFcov(dist=0, model=model, dim=2, seed=0)
  r <- RFgetModelInfo(RFcov, modelname="RMcutoff")$internalq[5] # theor R
  x <- seq(0, r, by= r / (size - 1)) * theta
  err <- try(RFsimulate(x, x, grid=TRUE, model=RPcirculant(model), n=0))
  return(if (class(err) == "try-error") NA else r)
}

alphas <- seq(1.52, 2.0, if (interactive()) 0.02 else 0.5) 
thetas <-
  if (interactive()) seq(0.05, 3.5, 0.05) else seq(0.1, 3.5, 2)

m <- matrix(NA, nrow=length(thetas), ncol=length(alphas))
for (it in 1:length(thetas)) {
  theta <- thetas[it]
  for (ia in 1:length(alphas)) {
  alpha <- alphas[ia]
  cat("alpha=", alpha, "theta=", theta,"")
  m[it, ia] <- stabletest(alpha=alpha, theta=theta)
  if (is.na(m[it, ia])) break
  }
  if (any(is.finite(m))) image(thetas, alphas, m, col=rainbow(100))
}

FinalizeExample()

Run the code above in your browser using DataLab