[R-sig-Geo] geographically weighted negative binomial regression model?

[Roger Bivand]

On Sun, 29 Oct 2006, Takatsugu Kobayashi wrote:

> Hi,
> 
> I wonder if anyone know whether there is a R package for implementing a 
> geographically weighted negative binomial regression model.

Not as far as I'm aware. There is draft ggwr() not yet released, which 
could possibly be extended, but you would be on untested ground - as is 
ggwr() anyway.

> Or is it possible to do this model by using a mgcv package with a family 
> = negative.binomial?
> I am not sure how to determine the adaptive bandwidth in a gam 
> package...... Does anyone out there kindly help?
> 

You could try fixing the other variables and letting gam() choose the 
"bandwidths" for s(x,y) by CV, if you like, but the "bandwidths" are not 
directly comparable with ggwr() bandwidths.

Roger

> 
> Thanks in advance.
> 
> Taka kobayashi
> Indiana Univeristy Bloomington
> PhD student
> 
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> 

-- 
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.Bivand at nhh.no