Dear Dave Miller,

I have no reference but a thought:

I think that a crossvalidation is unbiased (for a very complex error to be 
defined, which is however the same for both approaches) if the interpolation 
is really done only based on the choosen subsample. However depending on how 
we perform the crossvalidation there could be a problem with ordinary 
kriging:

If the variogramm is reestimated for each subsample, there should be no 
problem. But if we estimate the variogram once in the beginning and compare 
cross validation of an arbitrary preselected Radial Basis Function and a 
Ordinary kriging with an optimally fitted variogram to the complete dataset. 
I expect a small bias in favor of Ordinary kriging, because fitting a 
variogramm is already a kind of "finding the radial function, which will 
performs best with the given dataset". 

Best regards,
Gerald v..d Boogaart 


Am Mittwoch, 1. November 2006 10:51 schrieb Dave Miller:
> Dear Listers,
>
> When comparing cross-validated results between Ordinary Kriging and a
> variety of Radial Basis Functions, is there a reason that the RBF
> cross-validation statistics (e.g. ME, MAE, RMSE) may not be unbiased?
>
> Any references on this topic would be greatly appreciated.
>
> Regards,
>
> Dave
>
>
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Dave Miller
> GIS Research Assistant
> The Macaulay Institute
> Craigiebuckler
> Aberdeen
> AB15 8QH
>
> tel: +44 (0) 1224 498200 (switchboard) ext. 2261
> tel: +44 (0) 1224 498245 (direct dial option) ext. 2261
> fax +44 (0) 1224 311556
> e-mail: [EMAIL PROTECTED]
> websites: http://www.macaulay.ac.uk
> http://www.macaulay.ac.uk/LADSS/ladss.shtml

-- 
-------------------------------------------------
Prof. Dr. K. Gerald v.d. Boogaart
Professor als Juniorprofessor fuer Statistik
http://www.math-inf.uni-greifswald.de/statistik/  

B�ro: Franz-Mehring-Str. 48, 1.Etage rechts
e-mail: [EMAIL PROTECTED]
phone:  00+49 (0)3834/86-4621
fax:    00+49 (0)3834/86-4615   (Institut)

paper-mail:
Ernst-Moritz-Arndt-Universitaet Greifswald
Institut f�r Mathematik und Informatik
Jahnstr. 15a
17487 Greifswald
Germany
--------------------------------------------------

+
+ To post a message to the list, send it to [email protected]
+ To unsubscribe, send email to majordomo@ jrc.it with no subject and 
"unsubscribe ai-geostats" in the message body. DO NOT SEND 
Subscribe/Unsubscribe requests to the list
+ As a general service to list users, please remember to post a summary of any 
useful responses to your questions.
+ Support to the forum can be found at http://www.ai-geostats.org/

Reply via email to