Re: st: Adjusting Brian Poi's nlsurquaid.ado by removing symmetry conditions and `at' in -nl-

["Brian P. Poi" <brian@poiholdings.com>]

On 2/23/2011 3:10 PM, Alex Olssen wrote:
> Dear Brian,
>
> Thanks a lot for your response.  I managed to remove the symmetry conditions when doing it one at a time - I think I must have made a typing error somewhere.
> Your comment on `at' helps a lot.  As I understand it we need to send -nlsur- or -nl- a list of parameters to estimate.  If we don't name the parameters then we can just refer to them as elements of the vector `at'?  Renaming can make our code more readable, particularly when it contains many parameters.
>
> I think I'm going to try right up my own function evaluator from scratch and test it part by part.  Also, thanks drawing my attention to the variable() option.
>
> Kind regards,

When you use -nl- or -nlsur- with a function evaluator program, you 
specify the paramters by name in the parameters() option, or else you 
use the nparameters() option to indicate how many parameters there are.
The names you specify in the parameters() option control how the output 
is labeled, but as far as your function evaluator program is concerned, 
it just gets a vector named `at' that contains all the parameter values.

The temporary names you use inside your function evaluator program like 
g11, g21, etc. are there to help you, but as far as -nl- or -nlsur- is 
concerned, you could just as well refer to the elements of the `at' 
vector directly.  If you do go this route, I'd recommend using the same 
temporary names that you used in the parameters() option; otherwise 
you'll go crazy trying to figure out what is what.  But as far as the 
command itself is concerned, it couldn't care less whether you use 
temporary names at all.

  -- Brian Poi
  -- brian@poiholdings.com
*
*   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/