Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: pca and predict--confusion about what it does


From   Nick Cox <njcoxstata@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: pca and predict--confusion about what it does
Date   Sat, 20 Oct 2012 23:45:27 +0100

That's (what shall we say) one point of view. Another is that PCA is a
multivariate transformation technique. No model, no estimation. Factor
analysis is a branch of alchemy, so I'll refrain fom comment.

Also, the variances of the PCs are in general not equal, and so are
not entirely a matter of convention.

On Sat, Oct 20, 2012 at 11:35 PM, JVerkuilen (Gmail)
<jvverkuilen@gmail.com> wrote:
> On Sat, Oct 20, 2012 at 3:17 PM, Israel Pearce <ra.frbsf@gmail.com> wrote:
>> Thank you. I can't seem to find any options on Stata that do not scale
>> the PC's to have mean 0. Do you know of an option that could allow for
>> this or is it not a feature in Stata?
>
> The location and scale of the set of principal components is
> essentially arbitrary and hence fixed by convention. This isn't all
> that strange, given that the location of residuals in a regression are
> missing. PCA and factor analysis are bilinear regression models that
> make somewhat different assumptions but have the same identification
> issues.
*
*   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/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index