[R-sig-Geo] GWR and NAs

[Danlin Yu]

Raphael Saldanha:

Based on the output you provided, apparently these NAs are either 
ignored or treated as zeros. When you got a CV score as zero, you shall 
be aware that there is definitely a problem (all the adaptive qs produce 
the same 0, that means you are basically regression the dependent 
variable on itself, no independent variables participate in the 
regression (correct me if I am wrong).

I would suggest you investigate your data by exploring it with maps, and 
see if it is possible to ether interpolate those NAs or pick them out 
before you conduct a GWR analysis.

Hope this helps,

Danlin

Raphael Saldanha ??:
> Hi!
>
> I'm trying to use GWR with a dataset who has a lot of NAs. How spgwr
> considers NAs?
>
> As a test, I'm having the following results:
>
>   
>> g.adapt.gauss.00 <- gwr.sel(pib00 ~ cv00 + le00 + ma00 + ou00, data=dados,
>>     
> adapt=TRUE, coords=cbind(dados$x, dados$y))
> Adaptive q: 0.381966 CV score: 0
> Adaptive q: 0.618034 CV score: 0
> Adaptive q: 0.763932 CV score: 0
> Adaptive q: 0.854102 CV score: 0
> Adaptive q: 0.90983 CV score: 0
> Adaptive q: 0.944272 CV score: 0
> Adaptive q: 0.9655581 CV score: 0
> Adaptive q: 0.9787138 CV score: 0
> Adaptive q: 0.9868444 CV score: 0
> Adaptive q: 0.9918694 CV score: 0
> Adaptive q: 0.994975 CV score: 0
> Adaptive q: 0.9968944 CV score: 0
> Adaptive q: 0.9980806 CV score: 0
> Adaptive q: 0.9988138 CV score: 0
> Adaptive q: 0.9992669 CV score: 0
> Adaptive q: 0.9995469 CV score: 0
> Adaptive q: 0.99972 CV score: 0
> Adaptive q: 0.999827 CV score: 0
> Adaptive q: 0.999893 CV score: 0
> Adaptive q: 0.9999339 CV score: 0
> Adaptive q: 0.9999339 CV score: 0
> Houve 50 ou mais avisos (use warnings() para ver os primeiros 50)
>
>   
>> res.adpt.00 <- gwr(pib00 ~ cv00 + le00 + ma00 + ou00, data=dados,
>>     
> adapt=g.adapt.gauss.00, coords=cbind(dados$x, dados$y))
> Erro em gwr(pib00 ~ cv00 + le00 + ma00 + ou00, data = dados, adapt =
> g.adapt.gauss.00,  :
>   new data matrix rows mismatch
>
>
>
> Regards,
>
> Raphael Saldanha
> BRAZIL
> saldanha.plangeo at gmail.com
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>   

-- 
___________________________________________
Danlin Yu, Ph.D.
Assistant Professor of GIS and Urban Geography
Department of Earth & Environmental Studies
Montclair State University
Montclair, NJ, 07043
Tel: 973-655-4313
Fax: 973-655-4072
email: yud at mail.montclair.edu
webpage: csam.montclair.edu/~yu