[R] Model selection with BIC

[Thomas Lumley]

On Mon, 5 Mar 2001 s-luppescu at uchicago.edu wrote:

> Is there an efficient way to do linear model selection by choosing the model
> with the highest BIC from all possible models?

I wouldn't commit myself to this being an efficient way to do model
selection :), but the subsets() function in the leaps package does
all-subset regression, and the summary() function returns an object with
a $bic component.

Also, the plot method displays a summary of all the best models, showing
which variables are in each model. By default this is ordered by BIC. This
gives a slightly more representative view that just a best model.

There's no function to correct the bias in the coefficients of a model
chosen by all-subsets regression

	-thomas

Thomas Lumley			Asst. Professor, Biostatistics
tlumley at u.washington.edu	University of Washington, Seattle

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._