swh:1:snp:e520bf41b0e99213acde680a9d87fadac1aee079
Tip revision: da8174e204c4b3c8aff0fa179a0b53656129ef8e authored by Martin Schlather on 01 March 2004, 00:00:00 UTC
version 1.1.10
version 1.1.10
Tip revision: da8174e
RFtest.brown.space.time.R
# source("RFtest.brown.space.time.R")
###############################################################################
## Brownian motion, test
###############################################################################
if (file.exists("source.R")) source("source.R")
#if (file.exists("random")) load("random")
## 3d
RFparameters(pch="")
kappa <- 2 * runif(1)
scale <- 5 * runif(1)
var <- 5 * runif(1)
n <- 1000; TBM2.lines<-60
n <- 2; TBM2.lines<-2
#kappa <- scale <- var <- 1
x <- c(1,5,1)#x <- c(1,4,1)
z <- GaussRF(x=x, y=x, z=x, grid=TRUE, gridtriple=TRUE, model="3d",
me="local", param=c(0,var,0,scale,kappa), n=5 * n)
cat(Variogram(x=cbind(1,1,1), model="3d", param=c(0,var,0,scale,kappa)),
"",var(z[1,1,1, ]- z[2,2,2, ])/2,"\n")
cat(Variogram(x=cbind(1,1,0), model="3d", param=c(0,var,0,scale,kappa)),
"",var(z[1,1,1, ]- z[2,2,1,])/2,"\n")
cat(Variogram(x=cbind(0,0,1), model="3d", param=c(0,var,0,scale,kappa)),
"",var(z[1,1,1, ]- z[1,1,2,])/2,"\n")
cat(Variogram(x=cbind(1,3,4), model="3d", param=c(0,var,0,scale,kappa)),
"",var(z[1,1,1, ]- z[2,4,5,])/2,"\n")
# stop("")
print("A")
##2d
RFparameters(pch="")
kappa <- 2 * runif(1)
scale <- 5 * runif(1)
var <- 5 * runif(1)
#scale <- kappa <- var <- 1
x <- c(1,5,1)
z <- GaussRF(x=x, y=x, grid=TRUE, gridtriple=TRUE, model="2d",
me="local", param=c(0,var,0,scale,kappa) ,n=10 * n)
cat(Variogram(x=cbind(1,1), model="2d", param=c(0,var,0,scale,kappa)),
"",var(z[1,1, ]- z[2,2,])/2,"\n")
cat(Variogram(x=cbind(1,0), model="2d", param=c(0,var,0,scale,kappa)),
"",var(z[1,1, ]- z[2,1,])/2,"\n")
cat(Variogram(x=cbind(0,1), model="2d", param=c(0,var,0,scale,kappa)),
"",var(z[1,1, ]- z[1,2,])/2,"\n")
cat(Variogram(x=cbind(3,4), model="2d", param=c(0,var,0,scale,kappa)),
"",var(z[1,1, ]- z[4,5,])/2,"\n")
# stop("")
print("B")
DeleteRegister(0)
RFparameters(Print=7, TBMCE.trials=4, TBMCE.force=TRUE, pch="")
RFparameters(Print=4, TBMCE.force=TRUE, pch="#",
CE.mmin=8, TBMCE.strategy=1, TBMCE.trials=8,
TBMCE.mmin=c(-1,-3),
TBM2.linesimufactor=50,
TBM2.lines=TBM2.lines
)
######################################################################
## space-time
######################################################################
x <- c(1,2,1)
y <- c(4,5,1)
T <- c(1,2,1)
runif(1)
rs <- .Random.seed
#if (file.exists("random")) load("random") else
#{ runif(1); save(file="random",.Random.seed) }
var <- runif(1)
aniso <- cbind(matrix(runif(6),ncol=2), c(0,0,runif(1)))
#aniso <- diag(3) * 0.5
#aniso <- diag(2) * 10
kappa <- c(0.5, 2, runif(1,0,2), runif(1,0,1),
1 + 00000 * floor(runif(1,0,3)), 2 + rpois(1,2))
var <- 1 ###
model <- list(list(model="nsst", var=var, aniso=aniso, k=kappa))
#model <- list(list(model="sph", var=var, aniso=diag(c(runif(2), 0))),"*",
# list(model="exp", var=var, aniso=diag(c(0, 0, runif(1))))
# )
.Random.seed <- rs
print(model)
#x <- c(1,4,1)
print("C")
## stop("")
print(CovarianceFct(x=cbind(1,1,1), model=model))
print(Variogram(x=cbind(1,1,1), model=model))
z <- GaussRF(x=x, y=y, T=T, grid=TRUE, gridtriple=TRUE, model=model, me="TBM2",
n=n)
print(CovarianceFct(x=cbind(1,1,1), model=model))
print(cov(z[1,1,1, ], z[2,2,2,]))
print("")
print(CovarianceFct(x=cbind(1,0,0), model=model))
print(cov(z[1,1,1, ], z[2,1,1,]))
print("")
print(CovarianceFct(x=cbind(0,0,1), model=model))
print(cov(z[1,1,1, ], z[1,1,2,]))
print("D")
# stop("")
## does spherical work fine?
DeleteRegister(0)
RFparameters(Print=2, TBMCE.force=TRUE, pch="#",
TBMCE.strategy=1, TBMCE.trials=8,
TBMCE.mmin=c(-1,-3),
TBM2.linesimufactor=50,
CE.mmin=c(-1,-3),
CE.userfft=TRUE
)
aniso <- matrix(c(0.01,0,0,0.5),ncol=2)
model <- list(list(model="sph", var=var, aniso=aniso))
z <- GaussRF(x=c(1,75,1), y=y, grid=TRUE, gridtriple=TRUE, model=model, me="ci"
,n=3 * n)
print(CovarianceFct(x=cbind(1,1), model=model))
print(cov(z[1,1, ], z[2,2,]))
print("")
print(CovarianceFct(x=cbind(1,0), model=model))
print(cov(z[1,1, ], z[2,1,]))
print("")
print(CovarianceFct(x=cbind(0,1), model=model))
print(cov(z[1,1, ], z[1,2,]))
print(var(z))
print("D")
# stop("")
## this is used in combination with variable xx in RFcircembed.cc
## are all the elements of the FFT-matrix reached?
repeat {
nn <- 4 ## 1..4
x <- rbind(1,as.integer(runif(nn,2,2+rpois(1,100)^(1/nn))),1)
#x <- rbind(1,as.integer(runif(nn,2,2+rpois(1,10000)^(1/nn))),1)
print(x)
z <- GaussRF(x=x, grid=TRUE, gridtriple=TRUE, model="exp",
me="ci", param=c(0,1,0,0.0001) ,n=1)
cat("OK\n\n")
break; ##
}
# stop("")