Hoetker, Glenn <[email protected]> asks about -ml- programming:
> Hi all. I am trying to develop an ado file that calls a maximum
> likelihood estimator and am having a horrid time. I've included the
> code below. When it reaches " ml model lf palogit (`dep'=`xvars'
> `group') (delta: `group', nocons)", execution stops with "Unrecognized
> command: palogit
> program palogit not found".
> Yet from the documentation and various examples, defining palogit at the
> bottom of the program would appear to be a valid means of defining a
> sub-routine. What am I doing wrong? Do I need to have palogit defined
> as a separate ado file (I really hope not for ease of distribution)?
> (example code)
>
> program define complogit
> version 8.0
> syntax varlist, Group(varlist min=1 max=1)
> capture program drop palogit
>
> //Parse variable list
> gettoken dep xvars : varlist
>
> //Run logit equations and save results
> logit `dep' `xvars'
> estimates store all
>
> ml model lf palogit (`dep'=`xvars' `group') (delta: `group', nocons)
> ml search
> ml maximize
> end
>
> program palogit
> version 8
> args lnf theta delta
> quietly replace `lnf'= ///
> $ML_y1 * `theta' * (1+`delta')- ///
> ln(1+exp(`theta' * (1+`delta')))
> end
Glenn is correct that "sub-routine"s of a program in an ado-file are defined
in the same file after the main program definition. Such sub-routines are
then callable by the main program and other sub-routines but not by anyone
else.
Now the command (within -complogit-)
ml model lf palogit (`dep'=`xvars' `group') (delta: `group', nocons)
passes -palogit- as an argument to -ml model-. This "argument" is actually a
program name that -ml- will call to get likelihood values. Thus, -palogit-
must be contained in its own ado-file.
This is true of all commands that use -ml-.
--Jeff
[email protected]
*
* 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/
