James Hardy
>
> On 12/16/02 1:16 PM, "Nick Cox" <[email protected]>
> possibly may have written:
>
> > I don't know about partial least squares.
> > I dimly recall one user-programmer getting
> > interested in it some years ago, but I won't
> > name names just in case the project got
> > nowhere, or was private, or I'm mistaken.
> >
> > You get principal components from -factor-.
> > It would be easy enough to automate putting
> > them into a regression. On the other hand,
> > it may be that there is more to principal
> > components regression than calculating PCs
> > and using them as predictors in a regression,
> > which is what I'm assuming.
> >
>
> For PCR, the major difference would be having the ability
> to develop an
> optimum model using cross validation.
>
Yes; I wrote that with probability 1.
The cross-validation I think you would need
to write yourself.
Nick
[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/
