I ran into a new error and I cannot trace it because I think the problem
is caused by an ml proceedure.
I am running a ml model d0 estimation and this is my output:
initial: log likelihood = -1720.5831
alternative: log likelihood = -1685.3319
rescale: log likelihood = -801.15134
Iteration 0: log likelihood = -801.15134 (not concave)
Iteration 1: log likelihood = 193.03404 (not concave)
Iteration 2: log likelihood = 813.56394
Iteration 3: log likelihood = 974.59715
Iteration 4: log likelihood = 975.55101
Iteration 5: log likelihood = 975.55103
system limit exceeded - see manual
r(1000);
************************************************
The model is converging correctly because the last iteration
log-likelihood is the same as the true model from regress:
************************************************
. regress c y
Source | SS df MS Number of obs =
815
-------------+------------------------------ F( 1, 813)
= .
Model | 3249.51685 1 3249.51685 Prob > F =
0.0000
Residual | 4.35498339 813 .005356683 R-squared =
0.9987
-------------+------------------------------ Adj R-squared =
0.9987
Total | 3253.87184 814 3.99738555 Root MSE
= .07319
------------------------------------------------------------------------------
c | Coef. Std. Err. t P>|t| [95% Conf.
Interval]
-------------+----------------------------------------------------------------
y | .9982267 .0012816 778.86 0.000 .995711
1.000742
_cons | -.3667309 .0149193 -24.58 0.000 -.3960157
-.3374461
------------------------------------------------------------------------------
. di e(ll)
975.55103
Note this is a trivial example to replicate the error.
Any suggestions?
*
* 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/
