[R] Forward Selection with regsubsets

[Michael Dewey]

At 15:52 14/03/2008, wkoh at bigred.unl.edu wrote:
>Hi,
>
>I would like to perform a forward selection procedure on a data set
>with 6 observations and 10 predictors. I tried to run it with
>regsubsets (I set nvmax=number of observations) but I keep getting
>these warning messages:
>
>Warning messages:
>1: 5  linear dependencies found in: leaps.setup(x, y, wt = weights,
>nbest = nbest, nvmax = nvmax,
>2: nvmax reduced to  5 in: leaps.setup(x, y, wt = weights, nbest =
>nbest, nvmax = nvmax,
>
>I think the problem was due to inadequate observations, is there any
>way to perform subset selection in R if the number of observations are
>less than the predictors?

 > library(fortunes)
 > fortune("Yoda")

Evelyn Hall: I would like to know how (if) I can extract some of the
information from the summary of my nlme.
Simon Blomberg: This is R. There is no if. Only how.
    -- Evelyn Hall and Simon 'Yoda' Blomberg
       R-help (April 2005)

You have a number of options.
1 - Doing a site search for relaimpo would be a good start. Read the 
author's related articles first before trying to implement the procedures.
2 - Understanding your data better would help. Are you not curious as 
to what these dependencies are which R tells you about?
3 - Choosing predictors at random would be fast and in the absence of 
any clue about why you are doing this could be a good solution.

>Thanks!
>
>Woonyuen
>
>

Michael Dewey
http://www.aghmed.fsnet.co.uk