Statalist


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

RE: st: Re: computation of R-squared with a non-linear model


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   statalist@hsphsun2.harvard.edu
Subject   RE: st: Re: computation of R-squared with a non-linear model
Date   Wed, 20 May 2009 09:40:12 +0000 (GMT)

--- Marcel Spijkerman wrote:
> The output does not show any indication of a convergence
> problem. The model simply has a high R-square (I know that
> from previous research). But I am looking for something
> like the R-square of the untransformed model. 

Annything odd with -nl- automatically screams non-convergence,
and an R^2 of 1.000 certainly screams non-convergence to me. 
-nl- is very sensitive to correct starting values, as is
elaborately described in the documentation of -nl-. So I would
start there. Anyhow, I would always try multiple reasonable 
starting values with -nl-, just to check whether what I found
is really there, or just an artifact.

-- Maarten
 



      

*
*   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 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index