[R-sig-Geo] R-squared value for Spatial regressions

[Roger Bivand]

On Sat, 21 Feb 2009, Christina Qi Li wrote:

> Dear List members,
>
> It's a quite basic question, but I cannot figure it out myself: How to 
> get the R-squared value for a spatial lag regression, or a spatial error 
> regression?
>
> The default return for command "lagsarlm" or "errorsarlm" only returns 
> AIC value, not R-squared or Adjusted R-squared value.

Why basic? These are not defined for these kinds of models in the same way 
as for least squares, so use the AIC instead (uses both the log likelihood 
and adjusts for the number of included variables). You can fake R-squared 
via the correlation between the response and the fitted values, but it 
isn't what you expect.

Hope this helps,

Roger

>
> Anyone knows how to get it quickly? Thanks a lot!
>
> Christina
> University of Southern California
>
>
>
>
> 	[[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
>

-- 
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