Re: st: weights with Gaussian Kernel in lpoly, Stata 11

[Nick Cox <njcoxstata@gmail.com>]

I can reproduce this with Stata 12. I suggest contacting Stata tech support.

Nick

On Wed, Aug 29, 2012 at 10:18 AM, Nicolas Debarsy <ndebarsy@fundp.ac.be> wrote:
> Dear Statalist users,
>
> In a nonparametric regression using the lpoly command, I would like to put
> different weights on observations.
> However, when I want to use the Gaussian Kernel, I get the following error:
>
>              cross():  3200  conformability error
>   _lpoly_work():     -  function returned error
>             <istmt>:     -  function returned error
>
> For all the other kernels, the option [aw] runs.
> Could you tell me if there is something special with the Gaussian kernel or
> if it is a bug in the lpoly ?
> I'm using Stata 11.
>
> Here is the code I used:
>
> set obs 300
> drawnorm x3 e
> gen x32=x3^2
> gen y=x3+2*x32+e
> gen d=uniform()
> lpoly y x3 [aw=d], kernel(gaussian)
> lpoly y x3 [aw=d], kernel(epan2)
>
> Already many thanks for your answer.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/