Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Backed Up Iteration


From   Nick Cox <n.j.cox@durham.ac.uk>
To   "'statalist@hsphsun2.harvard.edu'" <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Backed Up Iteration
Date   Wed, 25 Jan 2012 20:57:53 +0000

Often and often. It is very difficult to tell from this kind of report whether you are trying to fit a model that is a bad idea for your data or the fitting process is just a bit tricky (or both). As you have several predictors, fitting an overcomplicated model really is a possibility, whatever the scientific (or non-scientific, e.g. economic) grounds for wanting to use them all. You could try tuning the -ml- engine by e.g. changing -technique()-. Simplifying the model first and then introducing complications gradually can sometimes isolate problematic predictors. 

Nick 
n.j.cox@durham.ac.uk 

Anna-Leigh Stone

I am using Stata 12.0 and I am attempting to run a fractional probit
regression with the command: glm dependent independent, fa(bin)
link(probit) cluster(gck). I have made sure that the dependent
variable values are not negative. I have 1500 dependent variable
observations out of 69,900 that fall at 0 and 1. Regardless of whether
I leave in the 0 and 1 values or take them out, I get the same log
likelihood iteration followed by backed up. It continues like this and
does not converge until I break it.
Has anyone else had this problem and know a solution to it? I do have
several variables but they are all necessary to my regression. I have
also tried the difficult option but that does not work either.

*
*   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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index