[rkward-devel] Many birds with the Zelig stone

[Thomas Friedrichsmeier]

Hi,

On Friday 25 September 2009, Michael Ash wrote:
> An easy way to substantially expand the functionality of rkward would
> be to enable access to the Zelig package.  Zelig is itself an umbrella
> interface to many of the statistical models in R.  See
> http://gking.harvard.edu/zelig/.  Instead of laboriously coding
> interface for lots of models (lm, glm, etc., etc.) rkward could simply
> code Zelig.

I admit I only took a *very* short look so far, and did not quite figure out 
what Zelig really does (lm, glm, lmer, and friends are already designed to 
have a very similar interface out of the box, and the differences in parameters 
and methods are mostly confined to the actual conceptual differences of the 
models. So if Zelig tries to abstract the differences, then what does it 
abstract?).

Well, I guess I'll need to take a better look, this was just my first glance. 
I'm just writing this mail to give you an answer to your question below, now, 
as I probably won't get around to writing mails on the weekend.

> I'd be pleased to work on the interface.  Where, btw, is the "simple
> regression" menu in rkward, which I hope to use as a template.

We have Analysis->Regression->Linear Regression 
(plugins/analysis/regression/linear_regression.*). If you go to Settings-
>Configure RKWard->Plugins and add the "under_development.pluginmap", you'll 
also get Analysis->Simple Anova (plugins/simple_anova/*). That is unfinished, 
but is likely interesting for its use of the <formula> element, which allows 
to construct basic interaction terms.

Regards
Thomas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/rkward-devel/attachments/20090925/9d31805c/attachment.sig>