This is a distinction without a difference. The
trick is merely to make sure that your own program
can pass options to -ml-. That will usually be
via a wildcard. After all, this is precisely what
many official commands do that get -ml- to do
the heavy lifting.
Nick
[email protected]
Feiveson, Alan
> for "official" commands, yes...but if you write your own -ml- program,
> -trace- comes as an option to -ml maximize- so in your -ml program you
> need something like
>
> ml maximize, trace
>
> If you want to have the trace turned on or off interactively, I think
> you need to include it as an argument or option in your ml program.
smerryman
> -trace-
>
> For example:
>
> . probit fore gear head, trace
>
> Iteration 0: log likelihood = -45.03321
> 0 0 -.5321897
> Iteration 1: log likelihood = -24.664128
> 1.907905 -.0665373 -6.200864
> Iteration 2: log likelihood = -21.824468
> 2.885617 -.1040464 -9.266334
> Iteration 3: log likelihood = -21.512585
> 3.320682 -.1300716 -10.62843
> Iteration 4: log likelihood = -21.506876
> 3.388441 -.1351156 -10.83891
> Iteration 5: log likelihood = -21.506874
> 3.389793 -.1352324 -10.84307
Rachel
> > Is there some option in ml that can make stata do this? I'm
> > interested in displaying the parameter vector b, not the predicted
> > values xb.
*
* 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/
