> > Is there anyone who can tell me
> > on which criterion is based
> > the 'regsubsets' function ?
> > Samuel


>the leaps help page says:
>
>cp or adjr2 or r2 is the value of the chosen model selection statistic for
>each model
>
>so each of these can be chosen as the selection statistic using the method
>argument
>
>hth, ingmar

In fact, 'regsubsets'
does not enable to choose
among cp, adrjr2 or r2
as 'leaps' function does.

Do you know the default criterion?
Or do you think it means that
the result would be the same
for the three criteria ?


Samuel BERTRAND
Doctorant
LBM - ENSAM - CNRS UMR 8005
http://bio-web.paris.ensam.fr

	[[alternative HTML version deleted]]