[R] step

[Uwe Ligges]

On 03.08.2011 16:33, Diana Schwegler wrote:
> Hello I am using the "step" function in order to do backward selection for a
> linear model of more than 200 variables but it doesn't work correctly.
> I think, there is a problem, if the matrix has same or more columns than
> rows.
> And if the matrix has too much columns the step-function doesn't work
> because the function will work with all columns together and I think, this
> is
> the problem.
> Is there a solution or a bug fixing of this problem?

If you reinvent the methodology.

Please read some textbook about linear models and how much sense a 
stepwise selection makes if you have so many variables that may be 
highly correlated etc....

Note that you do not have any degrees of freedom left for fitting the 
whole model. And that is required if you are aiming at a stepwise 
backward selection.

Uwe Ligges





> Thanks a lot
>
> --
> View this message in context: http://r.789695.n4.nabble.com/step-tp3715681p3715681.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.