Alfonso Miranda <[email protected]> posted that he was having some version
control issues with -ml-.
He sent me some files that represented the problem he was experiencing.
In Stata 11, -ml- was changed to be consistent with Mata's optimizers. This
means that method 'd2' evaluators are now expected to return a Hessian matrix
instead of the negative Hessian matrix.
However -ml- has the -negh- option which indicates that the method 'd2'
evaluator is going to return a negative Hessian matrix.
Those of you who want to call -ml- under -version 11-, but want to continue
using your existing method 'd2' evaluator without change, will need to specify
the -negh- option.
We should have emphasized this in -[U] 1.3 What's new-, but didn't.
--Jeff
[email protected]
*
* 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/
