Thomas,
You don't have to reinvent the wheel, why don't use -regress- or -_regress-
after this you can recover the parameter, type -ereturn list-.
For robust estimation use option -robust-, see the manual for formulae.
But if you still want to create a routine, take a look of -_robust- in the
Programming book.
Rodrigo.
----- Original Message -----
From: "Thomas Corneli�en" <[email protected]>
To: <[email protected]>
Sent: Wednesday, November 29, 2006 3:31 AM
Subject: st: Programming interface to reg ?
Dear Statalisters,
I wish to estimate the model y = X b + e by OLS.
I have a program that creates the matrices X'X and X'y of my problem in an
efficient way (loading the whole X matrix into the memory would be
inefficient because it is sparse).
My question is: After having constructed X'X and X'y (the system of normal
equations) can I provide these to Stata and tell it to do with it everything
it uses to do when it computes OLS estimates with the -reg- command? And
also use the options of -reg-?
If it is not possible at the step of X'X and X'y, then maybe can I use some
options of -reg- even after I have estimates of b and Var(b) at my disposal
? (My main concern is getting clustered or robust standard errors.)
Thanks for any suggestions!
Thomas
-------------------------------------------------------------------------
Thomas Cornelissen
Institute of Empirical Economic Research
Leibniz Universit�t Hannover, Germany
[email protected]
*
* 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/
*
* 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/
