Re: a principal component analysis question

Yiyu wrote:

No, the 2 groups can not be identified before the PCA was run. Actually
I hope to identify the 2 groups using the information from  the PCA.
What's more, even the number of groups can't be known before hand. ( K
can > 1000 and number of meaningful principal component can be > 5).

Please explain what you mean by "groups" and why you want to find them. Specifically, are "groups" combinations of variables, or combinations of observations/samples?

In your original post, "groups" seemed to be used to indicate variables (or combinations thereof). PCA, factor analysis and a few other related methods might work here depending on what you are trying to do.

In the usual way I think of "groups", this indicates a combination of observations which are similar; if that's what you want then you should be doing some type of cluster analysis.

--
Paige Miller
Eastman Kodak Company
paige dot miller at kodak dot com
http://www.kodak.com

"It's nothing until I call it!" -- Bill Klem, NL Umpire
"When you get the choice to sit it out or dance, I hope you dance" -- Lee Ann Womack
.

Relevant Pages

  • Re: TTest significance in replicate trials
    ... I will go out on a limb and say that without sufficiently detailed context, ... Eastman Kodak Company ... paige dot miller at kodak dot com http://www.kodak.com ... "When you get the choice to sit it out or dance, I hope you dance" -- Lee Ann Womack ...
    (sci.stat.edu)
  • Re: Mean and variance tests
    ... the conclusion of the mean test using the sample variance vars and the sample variance with a deleted value vars_x. ... Eastman Kodak Company ... paige dot miller at kodak dot com ... "When you get the choice to sit it out or dance, I hope you dance" -- Lee Ann Womack ...
    (sci.stat.edu)
  • Re: Mean and variance tests
    ... of location and spread that exist in the statistical literature ... Eastman Kodak Company ... paige dot miller at kodak dot com ... "When you get the choice to sit it out or dance, ...
    (sci.stat.edu)
  • Re: help with L9 array
    ... > Could we capture all main effects BETTER by running the L9 in its ... Eastman Kodak Company ... paige dot miller at kodak dot com ... "When you get the choice to sit it out or dance, ...
    (sci.stat.edu)
  • Re: Simple error analysis on slopes?
    ... Linear regression (see almost any stat text book) will compute the ... Eastman Kodak Company ... paige dot miller at kodak dot com ... "When you get the choice to sit it out or dance, ...
    (sci.stat.edu)
Loading