Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Steven Samuels <sjsamuels@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: RE: RE: RE: generalized nonlinear models |
Date | Mon, 8 Nov 2010 14:22:51 -0500 |
--Perhaps David could use -nl- after all. For computing the MLEs, -nl- lacks only the the ability to take an analytic weight that is a function of the parameters (Charnes, Frome, Yu (1976). The equivalence of generalized least squares and maximum likelihood in the exponential family, JASA 71:169-171. http://www.csm.ornl.gov/~frome/ JASAIWLS.pdf). One could add this ability in a program that alternates between an -nl- step with the -log4- option and a step that updates the analytic weight with the last parameter estimates.
Steve On Nov 8, 2010, at 1:27 PM, Feiveson, Alan H. (JSC-SK311) wrote:But -nl- optimizes by least squares, not by maximizing a Poisson (for example) likelihood. I think David's problem can be addressed by linearizing the response function in with respct to the parmameters and then use iterated Poisson regression. Of course to implement this, you'd have to calculate the first-order linear terms and also have a way to get first guesses of the nonlinear parameter(s).
Al -----Original Message-----From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu ] On Behalf Of Nick Cox
Sent: Monday, November 08, 2010 11:50 AM To: 'statalist@hsphsun2.harvard.edu' Subject: st: RE: RE: generalized nonlinear modelsUnless SAS and R have wired-in functions for your model, using -nl- in Stata strikes me as involving no more programming than using the equivalents in other languages.
Nick n.j.cox@durham.ac.uk Airey, David C Thanks, Stas and Al.Because the parameters of the 4 parameter logistic function used in dose-response analysis of drugs and chemicals on biological systems are the estimation goal, fitting that nonlinear function is desired.
This leaves me with these options SAS PROC NLMIXED R gnm package Programming Stata * * 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/ * * 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/ * * 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/