st: selecting the best combination of inputs

[Jeph Herrin <stata@spandrel.net>]

This is not really a Stata question, though if there is a Stata solution 
that would be excellent. Mostly, I am wondering if there is a name for 
this problem that would let me find more information on it.

I have a dependent variable -devpvar-, which is the outcome for a 
particular team; teams are hospital based, and may be comprised of one 
or more different medical specialties: physician, registered nurse, 
social worker, occupational therapist, etc. For each team I have a flag 
(-f1-...,-f10-) for each of the 10 specialties, indicating whether that 
team had an individual of that specialty.

What I would like to know is the best way to construct a team. I can 
easily examine how much each specialty contributes to the outcome using OLS

. reg depvar f1-f10

However, it seems like the only way to determine the optimum team using 
a similar model would be via including ~1k interaction terms  - I don't 
have to consider all possibilities, because there lower and upper limits 
on the team size, but this is still a bit much given that I have < 1000 
obs. Is there another method, unknown to me, for getting at an answer to 
this problem? If not a solution, is there at least name for this problem?

thanks,
Jeph
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/