[R] find bayesian information criterion for all variable combinations

[Ben Bolker]

eric lee <ericlee100 <at> gmail.com> writes:

> 
> Hi,
> 
> I have data for one dependent variable and multiple independent variables
> y = b0 + b1*x1 + b2*x2 + ...
> 
> I want to a list of all models that have some subset of the independents
> (just x1 x2, and not x3, etc.) and their corresponding BIC values.  Is there
> a pre-existing function that does this?  I saw that you can calculate
> individual BIC values using 'lm' and something like
> 
> AIC(lm1, k = log(nrow(swiss)))
> 
> but, nothing for all combinations.  Thanks.
> 
> eric

  There is a useful but rather dangerous package
available, not from CRAN, called "dRedging", that will
do this.

http://www.zbs.bialowieza.pl/users/kamil/r/

 USE AT YOUR OWN RISK -- misused, this will lead
you into terrible statistical practices.

  cheers
    Ben Bolker