I'm trying to work out a way to calculate anisotropy ratios in a simple case with the direction of the greatest and smallest continuity at right angles to each other. I thought that I would do this by minimizing the standard error of the variogram as in the following:
library(sp) data(meuse) coordinates(meuse) =~ x+y meuse = meuse["zinc"] meuse$zinc = log(meuse$zinc) data(meuse.grid) gridded(meuse.grid) =~ x+y library(automap) # available at http://intamap.geo.uu.nl/~paul/Downloads.html "f" <- function(ani, z, data) { [EMAIL PROTECTED], 2] <- [EMAIL PROTECTED], 2] * ani variogram = autofitVariogram(z, data) attr(variogram$var_model, "SSErr") } ani <- optimize(f, c(0, 100), maximum = FALSE, tol = 1, z = log(zinc)~1, data = meuse) meuseA <- meuse [EMAIL PROTECTED], 2] <- [EMAIL PROTECTED], 2] * ani$minimum kriging_resultA <- autoKrige(meuse, meuse.grid, meuseA) plot(kriging_resultA) # without anisotropy kriging_result = autoKrige(meuse, meuse.grid) plot(kriging_result) However, minimizing the standard error doesn't seem to be the right thing to do, as the larger I make the ratio, the smaller the standard error. Is there another measure I should be using?? Lisbeth, PS. Sorry, the meuse data set isn't the best example as it doesn't fit my assumption about anisotropy ratio at right angles, but everyone has access to it. Lisbeth R.i.i.s. S-c-a-r-a-b C-o-n-s-u-l-t ____________________________________________________________________________________ Luggage? GPS? Comic books? _______________________________________________ R-sig-Geo mailing list R-sig-Geo@stat.math.ethz.ch https://stat.ethz.ch/mailman/listinfo/r-sig-geo