Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Nick Cox <n.j.cox@durham.ac.uk> |
To | "'statalist@hsphsun2.harvard.edu'" <statalist@hsphsun2.harvard.edu> |
Subject | st: RE: non linear estimation |
Date | Wed, 23 Nov 2011 13:44:33 +0000 |
You can use whatever -nl- supports, which includes the specification of weights. But -nl- is not -regress- PLUS nonlinearity! If you want something beyond -nl-, you might need to do much more substantial programming, or approach it differently. This problem is fitting a power function with some constraints and -nl- is not the only way to do it. Nick n.j.cox@durham.ac.uk Graeml (Dainf) (a.k.a. Rodolfo) I´m using a substitutable expression programs (Nonlinear leastsquares estimation). The program (below) works very well. However, I would like to use robust option or some else that corrects heteroskedastic in nonlinear estimation. How can I implement it? Best regards, program define nlcobPR version 9.0 if "`1'" == "?" { global S_1 "a b c" global a=0.4 global b=0.4 global c=0.2 global S_2 "Cobb Douglas" exit } replace `1' = $a*((k^$b)*(l^$c)*(t^(1-$b-$c))) end nl cobPR y if uf=="PR" program drop nlcobPR * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/