Mohammed: I am sorry to burst your bubble, but the whole concept of finding a "best" subset regression is now considered to be badly flawed. There will typically be many essentially equivalent choices. Shrinkage of varying kinds and "ensemble fits" are the sorts of things that are done nowadays. I strongly recommend you work with a local statistician, as it is unlikely that you'll get the sort of help you need via the Net.
(All IMHO, of course). Cheers, Bert Bert Gunter Genentech Nonclinical Biostatistics (650) 467-7374 "Data is not information. Information is not knowledge. And knowledge is certainly not wisdom." H. Gilbert Welch On Sun, Mar 30, 2014 at 1:54 PM, Mohammad Ahmed <mwh...@yahoo.com> wrote: > I am not able to find the criterion under which the best subset model for a > certain size is chosen for the function regsubsets (leaps package in R) . For > leaps function, we can determine the method ( cp,r2 or adj r2). For example, > > x <- predictors > y <- response > > g <- regsubsets(x,y,nvmax=1,nbest=1) > > > summary(g)$which specifies set of predictors that form the best model within > all models of size 1. But under which criterion is considered best ?. Does it > have the highest R2 or adjR2 or any other criterion ?. > > > Thank you for your help. > > [[alternative HTML version deleted]] > > > ______________________________________________ > R-help@r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code. > ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
