[S] robust logistic regression

dunne@yarra.vut.edu.au
Wed, 14 Oct 1998 22:41:13 +1000


Dear S+ users,

has anyone implemented the robust logistic regression method of
Daryl Pregibon from

"Resistant Fits for Some Commonly Used Logistic Models with Medical
Applications, D. Pregibon, Biometrics, 38 (1982), pp. 485-498."

in S+?

If not -- what do people recommend for a robust logistic regression?

The robust(binomial) family does not seem to be robust enough as it
seems to taper the deviance function but does not downweight points
that are very influential due to their position in the feature space
(as D. Pregibon's method would do).

rob

-- 
* Rob Dunne  
* Victoria University of Technology , Footscray Campus      
* School of Communications and Informatics                
* P.O. Box 14428, MCMC,               Fax:   +61 3 9688 4050  
* MELBOURNE 8001, AUSTRALIA           Tel:   +61 3 9688 4757   
* http://matilda.vut.edu.au/~dunne  Email: dunne@matilda.vut.edu.au

Java has certainly revolutionized marketing and litigation. ----------------------------------------------------------------------- This message was distributed by s-news@wubios.wustl.edu. To unsubscribe send e-mail to s-news-request@wubios.wustl.edu with the BODY of the message: unsubscribe s-news