Re: st: Dropping variables in regression equations based on R-Squared value

[Nick Cox <njcoxstata@gmail.com>]

local droplist

foreach x of varlist   var1-var5000{
      quietly reg `x'  var5001
      display "`x'" `e(r2)'
      scalar a = e(r2)
      quietly reg `x'  var5002
      display "`x'" `e(r2)'
      scalar b = e(r2)
      if max(a, b) > 0.95 local droplist `droplist' `x'
}

drop `droplist'
Nick
njcoxstata@gmail.com


On 20 July 2013 07:45, George Murray <george.murray16@gmail.com> wrote:
> abcde
> Dear Statalist,
>
> I am running numerous regressions in the following form:
>
> foreach x of varlist   var1-var5000{
> quietly reg `x'  var5001
> display "`x'" `e(r2)'
> quietly reg `x'  var5002
> display "`x'" `e(r2)'
> }
>
> Where e(r2) displays the R-Squared of each regression.
>
> If *either* of the above regression equations have an R-Squared
> greater than 0.95, I want to drop the variable `x' . For example, if I
> regress var500 on var1001, and then on var1002, and *either* of the
> regressions have an R-Squared value over 0.95, I wish to delete
> var500.
>
> Does anyone have a way of doing this? I’m doing this manually at the
> moment, but is taking far too much time, and of course, it would be
> better if I could find a way to do this with the code.
>
> Thank you.
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/