Hi Edzer, 

I thought about that (the nugget effect) but my variogram (and robust
variograms) did show a short scale structure (with a high nugget
effect). 

Considering the possibility you mention, that I actually have a
phenomenon with a pure nugget effect although I see a structure, what do
I have to trust? My experimental variogram showing a structure or the
strong slope of my residuals (either + or -1 depending on the way you
define the residuals) plotted against the observed data telling me that
I actually may not have a structure? The correlation between my observed
and estimate values is poor indeed (0.40) but still.. 

For what concerns the definition of the residuals, Surfer and Isatis
used here defined the  residuals = estimate - observed. I haven't
checked the other software.

I had a quick look in Isaaks & Srivastava, Deutsch & Journel, Chiles and
Delfiner... all use the same definition as above (i.e. residuals =
estimate - observed). 

Many thanks again for your kind  feedback!

Gregoire

__________________________________________
Gregoire Dubois (Ph.D.)

European Commission (EC)
Joint Research Centre (DG JRC)
WWW: http://rem.jrc.cec.eu.int/

"The views expressed are purely those of the writer and may not in any
circumstances be regarded as stating an official position of the
European Commission."



-----Original Message-----
From: Edzer Pebesma [mailto:[EMAIL PROTECTED] 
Sent: 30 January 2008 14:13
To: Gregoire Dubois
Cc: [email protected]
Subject: Re: AI-GEOSTATS: Correlation between kriging residuals and
input data


Gregoire,

If you interpolate with a pure nugget effect, this is what you would 
expect for cross validation residuals because the predictions are 
constant, except that usually the residuals are defined as (observed - 
predicted) which would give the perfect correlation positive (1).

Which software gave you the residuals computed the other way around, or 
did you compute them yourself?
--
Edzer

Gregoire Dubois wrote:
>
> Dear list,
>
> Having fit a variogram to a dataset (indoor radon measurements) and
> applied cross-validations, I noticed the perfect negative correlation 
> (-0.95) between my kriging residuals and my input data.
>
> This means that I am overestimating as much the low values as I am
> underestimating the high values, something I am expecting since the 
> mean of the residuals  -> 0, a property of kriging. Fine so far.
>
> What I am puzzled about is of the possible reasons of getting such a
> strong slope (close to -1) of the plot of my residuals against my 
> input data?
>
> This, I understand, highlights that I am doing a systematic error
> somewhere which I want to avoid obviously. I thought I extracted 
> properly the spatially correlated component of my dataset (the 
> variogram of my residuals seems to show a pure nugget effect) but I 
> still can't find any reasonable explanation for the systematic errors.
>
> Any hints? I must have missed something obvious here.
>
> Many thanks for any feedback.
>
> Best regards,
>
> Gregoire
>

+
+ 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