• About DI
• Home
• Installation
• Manual/Tutorial   
• Example Dataset
• Related Software
• Fields
• Funfits

DI   se.image

Creates an image object of standard errors for spatial prediction 
using se.cov for the functions image or image.plot
 	
USAGE:
	se.image(nw.obj=new.network, nx=40, ny=40, ...)


REQUIRED ARGUMENTS:

nw.obj
	network object as created by make.network.obj

nx
	Number of x grid points

ny
	Number of y grid points

OPTIONAL ARGUMENTS:
...
	any other options for se.cov

VALUE:
	Returns a list of x, y, z where x and y are grid 
	locations and z is a matrix of standard errors of 
	prediction at each grid location.

SEE ALSO:
	se.cov

EXAMPLES:
# The function is currently defined as
function(nw.obj = new.network, nx = 40, ny = 40, ...)
{
	locs <- nw.obj$locs
	grid <- nw.obj$grid
	cov <- nw.obj$cov.obj
#
# discern what the sd object should be.
#
	if(is.null(cov$sd.obj))
		sdobj <- 1
	else if(is.numeric(cov$sd.obj))
		sdobj <- cov$sd.obj
	else sdobj <- get(cov$sd.obj)
	if(is.null(grid)) {
		x <- range(locs[, 1], na.rm = T)
		y <- range(locs[, 2], na.rm = T)
		x2 <- seq(x[1], x[2], length = nx)
		y2 <- seq(y[1], y[2], length = ny)
		grid <- list(x = x2, y = y2)
	}
	else {
		if(is.matrix(grid))
			grid <- list(x = grid[, 1], y = grid[, 2])
	}
	nx <- length(grid$x)
	ny <- length(grid$y)
	gridpts <- cbind(rep(grid$x, ny), rep(grid$y, rep(nx, ny)))
	cov.fcn <- cov$cov.function
	cov.args <- cov$args
	out <- se.cov(locs, gridpts, cov.function = cov.fcn, 
	       sd.obj = sdobj, cov.args)
	z <- matrix(out, nrow = nx, ncol = ny)
	list(x = grid$x, y = grid$y, z = z)
}