[R-sig-Geo] Recovering GWR Predicted Values from GGWR function

[Roger Bivand]

On Thu, 16 Jul 2015, Dexter Locke wrote:

> Hi list,
>
> Is there an in print citation to attribute this idea to "The aim of spgwr
> has always been to explore the weaknesses of GWR as a technique, not to
> help people use it"?
>
> Or should I use something like "(R. Bivand, personal communication)"?

Mailing lists are public communications, so cite by URL to the list 
archives if you so please.

Roger

>
> Thanks and happy R coding,
> Dexter
>
> On Wed, May 20, 2015 at 2:27 PM, Roger Bivand <Roger.Bivand at nhh.no> wrote:
>
>> On Wed, 20 May 2015, Brown, Jason wrote:
>>
>>  Hello,
>>>
>>> I'm estimating a Poisson GWR using the ggwr function.  I would like to
>>> recover the predicted values implied by the gwr model.  However, thus far
>>> I've only been able to recover the fitted values from the global model
>>> using $lm$fitted.values.  These obviously are the same values I get from a
>>> standard Poisson model. Is it possible to get the gwr predicted values
>>> instead?
>>>
>>
>> No, they are not provided. They would only be available if the data points
>> and the fit points were identical, or if newdata was available (and was
>> admitted by the function, which it is not). The aim of spgwr has always
>> been to explore the weaknesses of GWR as a technique, not to help people
>> use it, so no provision for GLM-GWR predictions is likely to be forthcoming
>> here (there are no coefficient standard errors either).
>>
>> Hope this clarifies,
>>
>> Roger
>>
>>
>>> Thank you,
>>> Jason
>>>
>>>
>>>
>>>         [[alternative HTML version deleted]]
>>>
>>> _______________________________________________
>>> R-sig-Geo mailing list
>>> R-sig-Geo at r-project.org
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>>
>>>
>> --
>> Roger Bivand
>> Department of Economics, Norwegian School of Economics,
>> Helleveien 30, N-5045 Bergen, Norway.
>> voice: +47 55 95 93 55; fax +47 55 95 91 00
>> e-mail: Roger.Bivand at nhh.no
>>
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> R-sig-Geo at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>

-- 
Roger Bivand
Department of Economics, Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; fax +47 55 95 91 00
e-mail: Roger.Bivand at nhh.no