plotGDdist: Scatterplot of sample spatial covariance entries and pairwise...

In guiludwig/bsplinedef: Spatial Deformation via Tensor Product of B-splines

Description

This function produces a plot of the upper diagonal sample covariance entries and the corresponding pairwise distances in the original and deformed coordinates, where the spatial deformation is obtained with a tensor product of B-spline basis functions, implemented in the bdef function.

Usage

plotGDdist(model, plot = TRUE, ggplot = FALSE, lines = TRUE)

Examples

set.seed(1)
n <- 40
x1 <- runif(n)
x2 <- runif(n)
x <- cbind(x1,x2)
F1 <- function(x1,x2) {
  x1 # (x1 - 2*x2)^2/sqrt(x1^2 + 4*x2^2)
}
F2 <- function(x1,x2) {
  sqrt(x1*x2) # (3*x1 + x2)^2/sqrt(9*x1^2 + x2^2)
}
plotGrid(list(window = list(x = range(x1), y = range(x2)),
              x = x, def.x = cbind(F1(x1,x2), F2(x1,x2))),
         F1 = F1, F2 = F2, margins = TRUE)
TIME <- 20
covModel2 <- RMexp(proj = "space", var = 1, scale = .25) *
  RMnugget(proj = "time", var = 1)
data2 <- RFsimulate(covModel2, x = F1(x1,x2), y = F2(x1,x2), T = 1:TIME)
y <- as.numeric(unlist(data2@data))
## Not run: 
matplot(1:TIME, t(matrix(y, ncol = TIME)), type = "b", lty = 1, pch = 1)

## End(Not run)
covModel2m <- RMexp(var = NA, scale = NA) # Only space?
testT <- bdef(x, y, tim = 1:TIME, cov.model = covModel2m, maxit = 10) # lambda estimated
plotGrid(testT, margins = TRUE)
plotGDdist(testT)

guiludwig/bsplinedef documentation built on May 16, 2020, 10:24 p.m.