RE: st: sb command + ereturn matrix

[michela bia <michela.bia@ceps.lu>]

Ok!
Thanks a lot Nick.

So, I cannot use the bootstrap in this way.

I will try to fix it.

Have a nice day,
Michela



-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Nick Cox
Sent: Thursday, January 12, 2012 2:18 PM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: sb command + ereturn matrix

I have very limited grasp of what you are trying to do, but as a
general strategic point it is difficult to see a case for ever wanting
to call -bootstrap- within a program (do-file, yes). -bootstrap-
potentially leaves lots of results, and then there is a collision
between its results and the results of your own program. Rather, have
your program save results, which you then -bootstrap- Of course you
can't -bootstrap- a matrix, which poses separate questions.

An even more general point is that it is best to write a program to do
one thing well, if only because a program that does everything you
want can be very much more difficult ever to get working.

More specifically, you write

"please note that they are saved in the local `DRFps' if I write mat
li `DRFps'!!)"

This is, or seems, very confused. -matrix list- just displays results,
and saves nothing. Also, the local macro referred to is just holding a
name for that matrix, and that local necessarily disappears at the end
of the program. Either way, there is no saving in that statement.

I can't follow the rest of your specific questions, but I suspect
-bootstrap- has just stomped on stuff you created earlier.

Nick

On Thu, Jan 12, 2012 at 12:47 PM, michela bia <michela.bia@ceps.lu> wrote:
> Dear all,
> I wrote an e-class program applying the radial penalized spline methodology and saving the estimates in a matrix (ereturn matrix command).
> I also made the bootstrap of standard errors, but the estimates don't seem to be saved in the matrix when I run the bs command
> (please note that they are saved in the local `DRFps' if I write mat li `DRFps'!!).
>
> The error message is: statistic matrix(r(DRFps)) evaluated to missing in full sample
> if I save the estimates writing : mat def DRFps = `DRFps'
>
> and it is: DRFps not found
> if I write:  ereturn matrix DRFps = `DRFps'
>
> May you please help me?
> How can I save the results in the correct way using the bootstrap at the end of the code?
>
> program define DRF, eclass
> version 11
>
> ...
> mat li `DRFps'
>
> _00006C[1,10]
>           c1         c2         c3         c4         c5         c6         c7         c8         c9
> r1  17349.756  13332.923  11062.302  9614.0808  8676.7353  8095.0012  7776.8476  7664.3185  7708.5554
>
>          c10
> r1  7883.4903
>
> ereturn matrix DRFps = `DRFps'
>
> if("`method'" == "radialpspline"){
>
> bs "DRF `varlist' `if' `in', outcome(`outcome') treatment(`treatment') gpscore(`gpscore') dose(`dose') method(`method') family(`family') link(`link') vce(`vce') common(`common') numoverlap(`numoverlap') tpoints(`tpoints') npoints(`npoints') npercentiles(`npercentiles') `derivative' nknots(`nknots') knots(`knots') `standardized' estopts(`estopts')" DRFps=matrix(r(DRFps)), nowarn `options'
>
> }
>
> End
>
>
> BTW - Reference:  'Semiparametric Regression', David Ruppert, M. P. Wand, R. J. Carroll July 2003

*
*   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/

*
*   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/