Thank you very much for your help.
However, I use -ml search- already.
In fact, the question arises when I have a constraint in my log-L function.
Initially, where is not constraint in my model, -ml maximize- does provide
estimated coefficient, though it takes a long long time to get converged.
But, when I set a constraint in one of variables, -ml maximize- can not do
anything because of failure to get a feasible initial value.
Does anyone have any idea?
Thank you very much
Jitian
----- Original Message -----
From: "Hung-Jen Wang" <[email protected]>
To: <[email protected]>
Sent: Monday, October 20, 2003 7:44 PM
Subject: Re: st: ML estimation
>
> Tuesday, October 21, 2003, 5:27:53 AM, Jitian Sheu wrote:
>
> > Dear listers:
>
> > I am doing MLE using command in Stata, -ml-
>
> > before I type: ml maximize, following the instruction of booklet of
"Maximum
> > Likelihood Estimation with Stata"(by William Gould and William Sribney).
> > I type: -ml check- first
>
> > The luck thing is that after -ml check-, the information I got is that
the
> > program I wrote has passed the all tests.
> > It seems that my estimation may go well.
>
> > however, when I type: -ml maximize-
>
> > the information I get is as follows:
>
> > initial: log likelihood = -2262.0106
> > rescale: log likelihood = -566.9123
> > rescale eq: log likelihood = -319.86063
> > initial values not feasible
> > r(1400);
>
> Try different initial values. You should find -ml plot- and -search-
> particularly helpful in this regard.
>
> HJW
>
> *
> * 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/
>
*
* 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/