On Aug 12, 4:43 pm, Paul.R.Sw...@UTH.TMC.EDU ("Swank, Paul R") wrote:
> Because they are perfectly inversely correlated and of course the eigenvalue is zero.
I still can't understand what you are saying. You make it sound like a
necessary condition for having zero variance has something to do with
the fact that the two values in the 4th eigenvector are equal ... or
that they are "inversely correlated" ... depending on which of your
posts I read.
I don't see this. I don't see any necessary or sufficient conditions
for having a zero variance for a component based upon the numbers
contained in the eigenvector.
paige\dot\miller \at\ kodak\dot\com