Dear list members
Standardized cross validation error is obtained by dividing the
difference between an observed value and its kriging prediction value by standard
deviation of corresponding kriging prediction.
It is expected that mean and variance of standardized cross validation
errors should be close to zero and one, respectively.
How close should those statistics be to zero and one?
Is it a usual way to test statistically these closeness?
Start your day with Yahoo! - make it your home page