Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: st: ml check with dfp technique
From
[email protected] (Jeff Pitblado, StataCorp LP)
To
[email protected]
Subject
Re: st: ml check with dfp technique
Date
Fri, 18 Jan 2013 11:20:08 -0600
Valerie Orozco <[email protected]> asks why -ml check- reports
an error message when used with -technique(dfp)-:
> I wonder why with Stata 12 it is no more possible to use" ml check" after "ml model" whereas it was possible with Stata 11 ?
> When I try to use it in Stata 12, I have the message :
>
> "technique(dfp) is not allowed with ml check
> r(198);"
The -ml check- command was designed to verify that the specified evaluator
function works properly with the given data, model specification, and evaluator
type. -ml check- performs the following sanity checks on the evaluator:
1. the evaluator does not modify the input vector
2. the evaluator will return the same non-missing criterion value upon
repeated calls with the same input vector
3. the evaluator returns consistent 1st derivatives when they are requested,
provided that 1st derivatives are expected with the given evaluator type
4. the evaluator returns consistent 2nd derivatives when they are requested,
provided that 2nd derivatives are expected with the given evaluator type
5. the evaluator doesn't produce unanticipated output
Some of these checks are irrelevant outside of -technique(nr)-; in fact, some
parts of -ml check- tacitly assume that -technique(nr)- is in control. For
this and other reasons, we changed -ml check- to report an error if anything
other than -technique(nr)- is being used. This change was made in the
18dec2012 update to Stata 12.
--Jeff
[email protected]
*
* 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/