Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: ml and non-linear coefficient


From   VIVIAND Xavier <Xavier.Viviand@mail.ap-hm.fr>
To   statalist@hsphsun2.harvard.edu
Subject   st: ml and non-linear coefficient
Date   Thu, 26 Sep 2002 11:46:32 +0200

Dear Statalist,


I posted a couple of weeks ago a message about fitting an Emax model (non
linear) with
several data from each subject. I was answered that the cluster option is
not allowed with the nl command and I should use ml command.
I am new to ml command and I am in trouble to fit a very simple non linear
equation :
	y = x^theta
I have simulated a data set of 100 observations:
	set obs 100
	gen x=-n
	gen y=x^3+10*invnorm(uniform())

Here is my do file :

capture program drop myproc
program define myproc
	version 7.0
	args lnf theta1 theta2
	quietly replace
`lnf'=ln(normd((y-x^`theta1')/`theta2'))-ln(`theta2')
end

ml model lf myproc (y,nocons ) /beta1 /sigma
ml search
ml maximize, difficult

I get  the following output :

. do essaiml
. capture program drop myproc
. program define myproc
  1.         version 7.0
  2.         args lnf theta1 theta2
  3.         quietly replace
`lnf'=ln(normd((y-x^`theta1')/`theta2'))-ln(`theta2')
  4. 
. end
. ml model lf myproc  (y,nocons ) /beta1 /sigma
. ml search
initial:       log likelihood =     -<inf>  (could not be evaluated)
could not find feasible values
r(491);
end of do-file
r(491);

Can somebody tell me what's wrong with my procedure.
Sincerely,

Xavier





*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index