Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: Re: RE: Fw: regarding stata MLE


From   Suddhasil Siddhanta <[email protected]>
To   [email protected]
Subject   Re: st: Re: RE: Fw: regarding stata MLE
Date   Thu, 15 Jun 2006 20:30:23 +0530

Dear dipankar

thank you very much.

the programme that you suggested is working but i am facing a problem when i put the command ml max, it returns error number 491.
the screen shows
initial: log likelihood = -<inf > (could not be evaluated)
could not find feasible values\
r(491);

I am using stata 7.0. special edition.

please suggest

best
suddhasil
----- Original Message ----- From: "Deepankar Basu" <[email protected]>
To: <[email protected]>
Sent: Thursday, June 15, 2006 8:03 PM
Subject: Re: st: Re: RE: Fw: regarding stata MLE



Suddhasil,

One problem that I see is that you have not specified an equation for
theta2 in your ml model command. You need to do that.

Second, when you run your program many times, you might encounter
problems if you do not use -capture program drop-; you need to add that
in the beginning of the program when debugging.

Third, since you are estimating the coefficients of a simple regression,
you might use the -normden- command in your likelihood evaluator. Your
likelihood functions seems to have missed the normal density altogether.
Here is what your lf program might be:

clear
capture program drop myprog
program define myprog
version 8.1 /* I assume you are using version 8.1 */
  args lnf theta1 theta2
  quietly replace 'lnf'= ln(normden($ML_y1, `theta1', `theta2'))
end

Then you can do the following from the command line:

.ml model lf myprog (reg: name of dependent variable = name of
independent variables) () /* The second bracket is for theta2 which will
give an estimate of the error variance. */

.ml check /* This will check your program for possible errors; if it
shows that your program has passed all tests, then go ahead and do the
next step. */

.ml max

Now, you should get the results.

HTH,

Deepankar Basu




On Thu, 2006-06-15 at 16:59 +0530, Suddhasil Siddhanta wrote:
Dear Nick,

I am now facing a problem in getting results for ml estimation in
stata.

the program that I have writen in a lf file is:-

program define myprog

args lnf theta1 theta2

quietly replace
'lnf'= -.5*ln(2*_pi)-.5*ln('theta2')-.5*(($ML_y1-'theta1')^2)/'theta2'

end

the statistical details that I have used to form the program is
1. the dependent vriable is normally distributed
the log likelihood is

]


then in the command window I typed

. ml model lf myprog (reg: name of dependent variable = name of independent
variables)

. ml maximize

now i am getting the eror 198 invalid name

please suggest me what i am doing wrong



Best,


suddhasil



----- Original Message ----- From: "Nick Cox" <[email protected]>
To: <[email protected]>
Sent: Thursday, June 08, 2006 6:29 PM
Subject: st: RE: Fw: regarding stata MLE


> This is already programmed for you as -regress-.
>
> Nick
> [email protected]
>
> N.B. Please do not send HTML to the list.
>
> Suddhasil Siddhanta
>
> I've just begun with the maximum likelihood with Stata.
>
> Can you please help me in this regard. My estimation equation is a > linear
> one with normal distribution.
>
> Y = constant + ax1 +b x2, where y is the dependent variable, a, b are > the
> coefficient and x1, x2 are the independent variables.
>
> Can you please help me in providing a programme that will run for the
> equation.
>
> *
> * 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/


*
*   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/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index