Dear STATAsticians,
In this day and age of plentiful computing power I favor the idea of
model optimization by just testing (an appropriate subset of) all
possible models (viz. combinations of independent predictors).
I have used the leaps-and-bounds Furnical & Wilson technique as
implemented in the vselect STATA program by Charles Lindsey & Simon
Sheather for linear regression.
Is there anything similar for use in the optimization of logistic
regression models?
Many thanks,
Nikos
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/faqs/resources/statalist-faq/
* http://www.ats.ucla.edu/stat/stata/
