st: function evaluator

[Peter Willemé <pw@plan.be>]

Dear all,
I keep getting the error (using Stata/SE 9.0)

(1997 missing values generated)
nllppl returned 199
verify that nllppl is a function evaluator program
r(199);

after running my program:

program nllppl
 syntax varlist (min = 2, max = 2) [if], at(name) lincoefs(numlist)
 tempname m1 w1 fi1
  local aa : word 1 of `lincoefs'
  local bb : word 2 of `lincoefs'
  local cc : word 3 of `lincoefs'
  local tau : word 4 of `lincoefs'
  scalar `m1' = `at'[1,1]
  scalar `w1' = `at'[1,2]
  scalar `fi1' = `at'[1,3]
 local logp : word 1 of `varlist'
 local t : word 2 of `varlist'
 replace `logp' = `aa' - `bb'*(`tau' - `t')^{m1} + `cc'*(`tau' - 
`t')^{m1}*cos({w1}*ln(`tau'-`t')-{fi1}) `if'
end


In short, this program should estimate 3 nonlinear parameters (m, w and 
fi) using -nl- :

nl lppl @ lnp t if tel < 61, parameters(m w fi) initial(m 0.59 w 1.83 fi 
175.3) lincoefs(`a' `b' `c' `tc')

There are 4 extra constants (a, b, c and tc) passed to the program, but 
they do not have to be estimated. What also puzzles me (and this may 
have something to do with the problem), is the statement "(1997 missing 
values generated)". I want to run the estimation on a subsample of the 
first 60 observations using -if-. This looks as if the sample selection 
does not work.

Many thanks for your help.
Peter




----------------------------------------------------------------------------

Disclaimer: please see "www.plan.be/disclaimer.html"

Please consider your environmental responsibility before printing this email

----------------------------------------------------------------------------

*
*   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/