Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Re: RE: vselect & ologit


From   Charles Lindsey <[email protected]>
To   [email protected]
Subject   st: Re: RE: vselect & ologit
Date   Thu, 10 Mar 2011 16:55:40 -0600

On Mar 10, 2011 Nick Cox <[email protected]> wrote:

> "-vselect- performs variable selection for linear regression."  
> As I understand it, there is no sense in which it supports
> -ologit- or works with -ologit-.  

> That would require major re-engineering of the program on your part 
> (indeed, I guess wildly, substantial innovation in terms of thinking of the 
> appropriate criteria to use).  

> Note: you say OLS, but -vselect- is not restricted to plain regression. 
> It can take various weights. Thus characterising it as using OLS is 
> incorrect.  Earlier today I commented on the bizarre practice of saying 
> "OLS" when regression in general is meant.  

> But I'd contact Charles Lindsey at StataCorp for the definitive
> word. (I've sent him a copy of this.)  

-vselect- does not yet support the general use of log-likelihoods,
particularly those corresponding to non-normal models.

An outline of how to perform variable selection for non-normal models is
given in:

Lawless, J. F. and K. Singhal 1978, Efficient Screening of nonnormal
        regression models.  Biometrics 34 318-327

We are currently working on implementing these methods into -vselect- for
submission to The Stata Journal.

Charles Lindsey
[email protected]

*
*   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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index