[R-sig-Geo] semivariogram of residuals from logistic model

[vieadmm4 at uib.es]

Dear all,

I wish to make a semivariogram using the residuals of a GLM  
(family=binomial(link=logit)). Well, my problem is that I have  
realized that residuals.glm() can return different types of residuals.  
I'm using the "response" type as it returns the values on the -1 to +1  
scale. However, I'm not quite familiar with this issue. Semivariograms  
of the different types look similar, but the values of semivariance  
are different.
I will appreciate any help regarding this issue and additional  
considerations that I should need to take into account for the  
variography of this kind of model.

Best regards,

David March Morla
Doctorando / PhD student
Instituto Mediterraneo de Estudios Avanzados (UIB-CSIC)
C/Miquel Marques,21
07190 Esporles
Islas Baleares - Spain
Tel: +34 971 611 722
Fax: +34 971 611 761
Email: david.march at uib.es