Thank you Roger.
I did'nt know what this LMZ.F3GWR.test() function was for. It seems that 
it is what I need, but when I try to run it, I have an error message 
coming up :

/Error: cannot allocate vector of size 52.5 Mo.

/I read on the R-sig-geo list that this could be due to the size of my 
dataset (3710 points) and that there was no way around... In that case, 
how to test big datasets ? The tests I'm doing now concern just one part 
of the total sample, so the problem will get bigger when I use all my data !
Would a random selection of points in the dataset be suitable for 
testing the spatial variability of the parameters ?

Thanks for your help

Marie

On Wed, 6 May 2009, Marie CHARRU wrote:

> Hi !
>
> I would like to know how to carry out the test for spatial variability 
> of the parameters of a geographically weighted regression model in R. 
> A test using a Monte Carlo approach is described in Fotheringham & 
> al., 2002, but I can't find how to run it with R.
> I would appreciate any help about this matter.

See the LMZ.F3GWR.test() function in the spgwr package for the Leung et 
al. test - other tests for GWR against OLS, but not for coefficients, 
are in that package. The help page has examples. Are you looking for 
other tests?

Hope this helps,

Roger

PS. Consider posting on the R-sig-geo list, where others may be able to 
help more.

>
> Thank you,
>
> Marie
> / /
>
>

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: roger.biv...@nhh.no

-- 
Marie CHARRU
Doctorante
UMR 1092 INRA/AgroParisTech - ENGREF
Laboratoire d'Etude des Ressources Forêt-Bois (LERFoB) - Equipe Ecologie 
Forestière
14, rue Girardet - CS 14216   -  54042 Nancy - France
Tel. : 03.83.39.68.80  - Fax. : 03.83.30.22.54


        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
R-sig-Geo@stat.math.ethz.ch
https://stat.ethz.ch/mailman/listinfo/r-sig-geo

Reply via email to