Hi
I have a problem using the ml init command
I want to estimate a model with a restriction one a coefficient not being negative, using a procedure I found on http://www.stata.com/support/faqs/stat/intconst.html.
But i run into trouble during the 2nd estimation (setting separate equations for the coefficients) with a log-likelihood = -inf
I thought of using the residuals of an intial regression as starting values for sigma, but I cannot figure out how to use the ml init command. Can anyone help?
My program looks like this;
/* begin do file */
cscript
program mynormal_d0
version 9.2
args todo b lnf
tempname lnsigma sigma
tempvar mu
mleval `mu' = `b', eq(1)
mleval `lnsigma' = `b', eq(2) scalar
quietly {
scalar `sigma' = exp(`lnsigma')
mlsum `lnf' = ln( normalden($ML_y1,`mu',`sigma') )
}
end
use "D:\Stud.ba.oecon\Bachelor Opgave\STATA\datasaetML"
ml model d0 mynormal_d0 (xb: return_rf = divdum beta d_yield inst_ mv__mil__) (lnsigma:), ///
diparm(lnsigma, exp label(sigma))
ml maximize
/* end do file */
/* setting separate equations for the coefficients */
/* begin do file */
cscript
program mynormal_b
version 9.2
args todo b lnf
tempname a lnsigma sigma
tempvar xb mu
mleval `a' = `b', eq(1) scalar
mleval `xb' = `b', eq(2)
mleval `lnsigma' = `b', eq(3) scalar
quietly {
generate double `mu' = `xb' +`a'*$x2
scalar `sigma' = exp(`lnsigma')
mlsum `lnf' = ln( normalden($ML_y1,`mu',`sigma') )
}
end
use "D:\Stud.ba.oecon\Bachelor Opgave\STATA\datasaetML"
global x2 beta
ml model d0 mynormal_b (a: ) (xb: return_rf = divdum d_yield inst_ mv__mil__) (lnsigma:), ///
diparm(lnsigma, exp label(sigma))
ml maximize
/* end do file */
I thank you for your interest
Best regards
Martin Heissel
stud.bach.oecon
+4520120025
Thorvaldsensgade 11, 1. TH
DK-8000 �rhus C
*
* 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/