[R-sig-Geo] coff standard erros in ggwr

[Roger Bivand]

On Tue, 16 Sep 2008, Nelson Oliveira wrote:

> I'm using the function ggwr (for the first time). I want to test the
> significance of the coeff. estimates (global), but I can't find their
> standard errors.  I used the following code:
> xx<-SpatialPointsDataFrame(coords=coord,data=en1)
>
> bw <- ggwr.sel(numhomicidio~offset(log(popmedia))+I12 + I4 + I5 +
> I10,data=xx,family=quasipoisson())
>
> ed_quasi<-ggwr(numhomicidio~offset(log(popmedia))+I12 + I4 + I5 + I10,
> data=xx,family=quasipoisson(),bandwidth=bw)

The local coefficient standard error estimates are not calculated or 
returned by ggwr(). They could in principle be computed, but would you 
trust them? Since gwr() and ggwr() are seriously affected by induced 
collinearity, I doubt whether inference is where one wants to go. Please 
do try:

pairs(as(ed_quasi$SDF, "data.frame")[, 2:5])

for a quick check (Wheeler & Tiefelsdorf, 2005).

Roger

>
> Thanks
> Nelson
-- 
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