Date: Wed, 12 Aug 2009 15:43:15 -0500 "Swank, Paul R" "SAS(r) Discussion" "Swank, Paul R" Re: Principal Component Analysis To: Paige Miller text/plain; charset="us-ascii"

Because they are perfectly inversely correlated and of course the eigenvalue is zero.

Dr. Paul R. Swank, Professor and Director of Research Children's Learning Institute University of Texas Health Science Center-Houston

-----Original Message----- From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Paige Miller Sent: Wednesday, August 12, 2009 2:44 PM To: SAS-L@LISTSERV.UGA.EDU Subject: Re: Principal Component Analysis

On Aug 12, 12:27 pm, Paul.R.Sw...@UTH.TMC.EDU ("Swank, Paul R") wrote: > The eigenvalue is the variance of the weighted component so it reflects all the measure that are included in the component. Note, for example, that S2 and S4 are perfectly inversely related. Thus the eigenvectors for the fourth component are equal and positive for S2 and S4 which means the variance of the component will be zero.

I'm afraid I can't make any sense out of this last sentence. Two values of a 4-dimensional eigenvector are equal and positive (I agree so far); how does that imply the variance of the component (eigenvalue) is zero?

-- Paige Miller paige\dot\miller \at\ kodak\dot\com

Back to: Top of message | Previous page | Main SAS-L page