# Re: st: sorted rolling regresion

 From "William Bishop" To statalist@hsphsun2.harvard.edu Subject Re: st: sorted rolling regresion Date Mon, 4 Feb 2008 15:50:26 -0800

```Scott and Austin:

Thank you both for your help with the rolling regression.  That did
what I was hoping to accomplish.

BTW, when I modified Scott's code to be time<`i' instead of time<=`i'
it stopped making predictions.  I assume that has to do with the
e(sample) line,  which I am unfamiliar with - will have to look into
that.

Thanks again.

William

On Feb 4, 2008 1:21 PM, Austin Nichols <austinnichols@gmail.com> wrote:
> William Bishop <wbishopco@gmail.com>:
> Scott's code does not make my mistake of leaving a condition off -predict-:
>
> webuse grunfeld, clear
> su time, meanonly
> loc Max=r(max)
> levelsof com, local(is)
> g p1=.
> foreach i of local is {
>  forval t = 3/`Max' {
>  qui reg mval kst if com==`i'&time<`t'
>  predict temp, xb
>  qui replace p1=temp if com==`i'&time==`t'
>  drop temp
>  }
> }
> li time p1 mval kst, sepby(com) noo
>
> but note (still) that the predictions may make no sense.  You may
> prefer a different flavor of  regression, depending on what you are
> going to use this for...
>
> On Feb 4, 2008 4:06 PM, Scott Merryman <scott.merryman@gmail.com> wrote:
> > On Feb 2, 2008 1:47 PM, William Bishop <wbishopco@gmail.com> wrote:
> > > I'm trying to do a rolling regression by id that regresses on previous
> > > dates and then gets the prediction for the current observation.
> >
> > No, but you can loop over the groups.  Perhaps, something like this will help:
>
> *
> *   For searches and help try:
> *   http://www.stata.com/support/faqs/res/findit.html
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
```