Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: SE matsize 11000


From   Ulrich Kohler <kohler@wz-berlin.de>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: SE matsize 11000
Date   Tue, 30 Aug 2005 17:19:39 +0200

Julia Nasev wrote:
> my question is related to the bottleneck of SE being a matsize of
> 11000. I understand that the matrix size is dependent on the
> estimator you employ but I don't yet have a feeling how the matrix
> size and the number of observations (e.g. the number of companies)
> are related. We don't know yet which models we will be using but
> let's say that our dataset consists of about 2000 companies/15 years
> (unbalanced panel) and 10 independent variables. Assume we want to
> apply linear (and/or dynamic) panel models. Besides the OLS estimator
> we will need GSL/EGSL, GMM estimators etc. How can I get a feeling of
> the max. number of companies/years I can work with in Stata?

The number of observations is almost completely unrelated to the matsize. What 
counts are the number of independent variables in your statistical models. 
You will rarely have more than 11000 independend variables in the model.  
Therefore I would not regard the matsize of 11000 as a bottleneck. Even the 
800 maximum matsize in Intercooled Stata is not very restrictive in this 
respect.

Matsize can sometimes be harmful in the context of multivariate statistics 
and/or cluster-analysis. In such cases Mata provides a way out. 

Uli

-- 
kohler@wz-berlin.de
+49 (030) 25491-361
*
*   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/



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