Message-ID: <Pine.LNX.4.43.0902220352570.31465@hymn11.u.washington.edu>
Date: 2009-02-22T11:52:57Z
From: Thomas Lumley
Subject: variable/model selction (step/stepAIC) for biglm ?
In-Reply-To: <440d7af40902220349m58a0f897ja59457b34c7ef7a7@mail.gmail.com>
On Sun, 22 Feb 2009, Tal Galili wrote:
> Hi Thomas,
>
> What you just wrote is very interesting to me - do you have any suggestion
> then as to how to implement leaps (or any other package/code) to iterate on
> the final lm model produced by biglm ?
> Any advice would be very welcomed!
>
If you look at the source code in the leaps package, it first sets up a QR decomposition and then calls search routines. The QR decomposition code is exactly the same Fortran code as is used by biglm, so you should be able to plug the output of biglm into the subset selection code.
-thomas
Thomas Lumley Assoc. Professor, Biostatistics
tlumley at u.washington.edu University of Washington, Seattle