R help - Aug 2001 - variable selection in R using bootstrapping

Hi, all:

Is there anyone know about the variable selection procedure in R using
bootstrapping? I tried to define the statistics function as following to find
the best set of covariates in each bootstrap resample. My plan was to construct
the final model based on the common covariates picked from every R=300
resamples.
 
dat3.pred.fun <- function( data, i, formula )

{ d <- data[i,]

  d.glm.all <- glm(formula,family=binomial,data=d )

  d.fit<-step(d.glm.all)

  formula(d.fit) 

# or summary(d.fit)$coefficient[,1]

# .....

}

 And then I called the boot function as: 

r_300  # no of bootstrap samples

dat3.boot <- boot( dat3, dat3.pred.fun, R=r,sim="ordinary",
formula=formula(dat3.glm.help))

However, R gave me message like " terms. formula(X, ...): object
"d" not found". Anyone has any clue about it?

Thank you very much for your time!

Sincerely,

Yunda

Department of Biostatistics 

University of California, Los Angeles








-------------- next part --------------
An HTML attachment was scrubbed...
URL:
https://stat.ethz.ch/pipermail/r-help/attachments/20010820/8be2d322/attachment.html

Please use ordinary text e-mail.

Bootstrapping has no advantage for selecting variables,
only for studying the ill effects of such selection.
The major problem with using the bootstrap the way
you have outlined is that the selection frequency is
ruined by collinearity, i.e., collinearity makes the
selection of one variable over another about as
reliable as flipping a coin.  Besides, the selection frequency
is highly related to the P-value from the initial model,
if you were using backwards stepdown.  So the bootstrap
does not offer much new information anyway.  -Frank
-- 
Frank E Harrell Jr              Prof. of Biostatistics & Statistics
Div. of Biostatistics & Epidem. Dept. of Health Evaluation Sciences
U. Virginia School of Medicine  http://hesweb1.med.virginia.edu/biostat
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._