gstat / Linear model of coregionalization

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

gstat / Linear model of coregionalization

Nick Hamm-2
Dear all
Is anybody using gstat for computing cross-variograms and fitting cross variogram models?  I have implemented this in a similar manner to the example code given in the gstat vignette (Section 10).

I would like to extract the coefficients (i.g., the  b_{ij}^l of the linear model of coregionalization as given in Equation 4.37 of Goovaearts (1997) book (page 110). 

\gamma_{ij} (h) = \sum_{l=0}^L b_{ij}^l g_l(h) \forall i, j 

image.png



Is this possible?

kind regards

Nick



_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Reply | Threaded
Open this post in threaded view
|

Re: gstat / Linear model of coregionalization

Manuel Ribeiro

Hello Nick,

 

Once you have your lmc fitted, call the object $psill.

For example, if your lmc object is called g.fit with variables v , z call:

 

g.fit$model$v$psill[1] to get coefficient of v for 1st structure of the nested model.

g.fit$model$z$psill[1] to get coefficient of z for 1st structure of the nested model.

 

g.fit$model$v$psill[2] to get coefficient of v for 2nd structure of the nested model.

g.fit$model$z$psill[2] to get coefficient of z for 2nd structure of the nested model.

 

For cross-variogram, please see the list of g.fit$model to see the name of the object storing v*z cross-variogram model coefficients.

 

I hope this helps. There are other ways of doing this, anyway this one should work.

Regards,

Manuel

 

 

Manuel Luís Castro Ribeiro

 

CERENA- Centro de Recursos Naturais e Ambiente


Instituto Superior Técnico
Av. Rovisco Pais 

1049-001 Lisboa

Portugal


Phone:  +351 21 841 74 41

Mobile: +351 92 442 80 07

 

Web: www.cerena.pt

 

 

 

From: R-sig-Geo <[hidden email]> On Behalf Of Nick Hamm
Sent: quarta-feira, 29 de janeiro de 2020 04:22
To: r-sig-geo <[hidden email]>; [hidden email]
Subject: [R-sig-Geo] gstat / Linear model of coregionalization

 

Dear all

Is anybody using gstat for computing cross-variograms and fitting cross variogram models?  I have implemented this in a similar manner to the example code given in the gstat vignette (Section 10).

 

I would like to extract the coefficients (i.g., the  b_{ij}^l of the linear model of coregionalization as given in Equation 4.37 of Goovaearts (1997) book (page 110). 

 

\gamma_{ij} (h) = \sum_{l=0}^L b_{ij}^l g_l(h) \forall i, j 

 

image.png

 

 

Is this possible?

 

kind regards

 

Nick

 

 


_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo