[R-sig-Geo] How to check model assumptions of a weighted spdep::spautolm?

[Roger Bivand]

On Thu, 2 Jul 2015, Guido Schulz wrote:

> I would like to check the model assumptions of a weighted
> spdep::spautolm model.
>
> i.e., I would like to check if the (marginal) residuals, defined as Y -
> (Signal Trend + Signal Stochastic) are homoscedastic, independent and
> normal.
>
> Unfortunately spdep::bptest.sarlm does not accept spautolm objects,
> hence I cannot carry out a Breusch-Pagan test - unless there is a
> workaround?
>
> The only ideas I had so far:
>
> - qqnorm & qqline plot of the residuals
> - visual check of a map of the residuals (to identify obvious remaining
> heteroscedasticity or autocorrelation)

This is correct. The adaptation in spdep::bptest.sarlm is also untested, 
and was based on what was known when that was written. The underlying 
problem is that heteroscedasticity and spatial autocorrelation can be 
difficult to separate.

Hope this helps,

Roger

>
> Any suggestions?
>
> Best,
>
> Guido
>
> _______________________________________________
> 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