st: RE: Stata 11 discrim logistic leave-one-out

["Nick Cox" <n.j.cox@durham.ac.uk>]

You asked the same question on 31 March. 

<http://www.hsph.harvard.edu/cgi-bin/lwgate/STATALIST/archives/statalist
.1003/date/article-1976.html> 

Evidently, no one has an answer. I certainly don't. 

I think your best option is to ask Stata tech support. 

Nick 
n.j.cox@durham.ac.uk 

Burkhard C. Schipper

How to do leave-one-out classification tables for logistic
discriminant analysis?

I am using discrim command for discrimination analysis in Stata 11.
For linear, quadratic, or knn discriminant analysis I can produce
leave-one-out classification tables by achieve this by typing
respectively

discrim lda ... , ... nootable lootable
discrim qda ..., ... nootable lootable
discrim knn ...., ... nootable lootable

or by using

estat classtable, loo

after performing the discrim commands respectively. This is discussed
nicely in the Stata 11 manual.

However, for logistic discriminant analysis, the Stata 11 manual does
not discuss lootable or loo. In fact, if one types

discrim logistic ...., lootable notable

one obtains an error message

option lootable not allowed
r(198);

A similar error message occurs after

discrim logistic ..., ...
estat classtable, loo

Why is leave-one-out not allowed in logistic discriminant analysis?

Does anyone know a way around this problem?

*
*   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/