No doubt this could continue for some while. In my case, the need to
sign off dictates my last comment. The lack of an error model for a PCA
need worry no one any more than the lack of a model for logarithmic
transformation! Sometimes it's helpful; sometimes it is not.
Nick
[email protected]
Verkuilen, Jay
Nick Cox wrote:
>>I'd go a little further. There is no model underlying PCA. It's a
transformation procedure!
Yes, I could see that. It certainly is a very useful pre-whitener or,
conversely, low-pass filter, for multivariate data.
Lack of an error model is, IMO, a big issue.
*
* 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/
