| |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: Streg, Distri(logn) convergence issues
--- Shon Hiatt <[email protected]> wrote:
> I am using a natural logarithm event history analysis [streg,
> distribution (Logn)] and I continually obtain the following errors
> after multiple iterations:
>
> Discontinous region encountered
> Cannot compute an improvement
>
> Should I change the convergence tolerances (as shown under the Stata
> help "Maximize")? Any suggestions would be appreciated.
No, it probably means that there is something wrong with your model
(the log normal distribution is not appropriate, you have variables
that are (almost) collinear, etc.) If you are really sure there is
nothing wrong with your model I would look for better starting values.
Try for instance to estimate an -streg- with the loglogistic
distribution (which is pretty similar) and use those parameters as
starting values. But if the loglogistic converges on its own and the
lognormal doesn't than that might tell you something...
Hope this helps,
Maarten
-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands
visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434
+31 20 5986715
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
___________________________________________________________
New Yahoo! Mail is the ultimate force in competitive emailing. Find out more at the Yahoo! Mail Championships. Plus: play games and win prizes.
http://uk.rd.yahoo.com/evt=44106/*http://mail.yahoo.net/uk
*
* 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/