Hi everyone,
I am working on rainfall interpolation using regression kriging method and I
need suggestions on how I can carry out a cross validation (leave-one-out)
for this elaboration. At first I tried to apply directly Krige.cv, similarly
to UK method (example for october: PP10uk.cv -
Hello Michele,
Universal kriging is equivalent to Linear Regression (with the
generalized-least-squaresestimator) + Simple Kriging of residuals (e.g.
Cressie, 1993, section 3.4.5). The differences you observe are probably
due to the use of ordinary least squares. If you use