Author 
Message 
JEMebius science forum Guru Wannabe
Joined: 24 Mar 2005
Posts: 209

Posted: Mon Jul 17, 2006 1:46 pm Post subject:
Re: symmetry in eigenvectors



Vector y=(x2 ,x1 ,x4 ,x3) is obtained from vector x=(x1, x2, x3, x4)
by a double 4D rotation over 90 deg in the 12 and the 34 planes.
Whereas it is true that the other three eigenvectors of A are orthogonal
to x, there is no a priori reason to expect that y =must= be one of
these three.
However, by a suitable 4D coordinate transformation one can arrange
things such in the new coordinates x'1, x'2, x'3, x'4 the vector x is
unchanged: x = (x'1, x'2, x'3, x'4), and one of the remaining three
eigenvectors becomes (x'2, x'1, x'4, x'3).
It is a worthwhile exercise to find out which 4x4 symmetric matrices
have (x1, x2, x3, x4) and (x2, x1, x4, x3) as two of their eigenvectors.
Ciao: Johan E. Mebius
Al wrote:
Quote:  Hi!
Maybe someone could help me with the following. It would be very
helpful!
Given a real symmetric 4x4 matrix A (it's a covariance matrix), and one
real eigenvector x=(x1,x2,x3,x4).
Is it now possible to show that y=(x2,x1,x4,x3) is an eigenvector of
A, too?
Thanks,
Al



Back to top 


Chip Eastham science forum Guru
Joined: 01 May 2005
Posts: 412

Posted: Mon Jul 17, 2006 1:32 pm Post subject:
Re: symmetry in eigenvectors



Al wrote:
Quote:  Hi!
Maybe someone could help me with the following. It would be very
helpful!
Given a real symmetric 4x4 matrix A (it's a covariance matrix), and one
real eigenvector x=(x1,x2,x3,x4).
Is it now possible to show that y=(x2,x1,x4,x3) is an eigenvector of
A, too?

No, it isn't. While y is orthogonal to x, it is not true that
just because A is real symmetric, and x is an eigenvector
that y must also be an eigenvector.
It is true that eigenvectors of real symmetric A for distinct
eigenvalues must be orthogonal.
Here's a way to construct counterexamples to your
underlying claim. Take x (normalized) and extend
it to an orthonormal basis {x, u, v, w} for R^4 such
that y is not a multiple of any single basis element.
Define A = x'x + 2 u'u + 3 v'v + 4 w'w.
Clearly A is real symmetric. Due to orthonormality
of the basis, the eigenvalues are shown to be 1,2,3,4
corresponding to resp. basis x,u,v,w of eigenvectors.
Since y is not a multiple of any of these, y is not an
eigenvector of A.
regards, chip 

Back to top 


A N Niel science forum Guru
Joined: 28 Apr 2005
Posts: 475

Posted: Mon Jul 17, 2006 1:16 pm Post subject:
Re: symmetry in eigenvectors



In article <1153141791.978301.175850@s13g2000cwa.googlegroups.com>, Al
<alalcoolj@gmx.de> wrote:
Quote:  Hi!
Maybe someone could help me with the following. It would be very
helpful!
Given a real symmetric 4x4 matrix A (it's a covariance matrix), and one
real eigenvector x=(x1,x2,x3,x4).
Is it now possible to show that y=(x2,x1,x4,x3) is an eigenvector of
A, too?

No.
Say A is diagonal matrix 1,1,1,0. Then (1,0,1,0) is an eigenvector,
but (0,1,0,1) is not. 

Back to top 


Al1131 science forum beginner
Joined: 01 Mar 2006
Posts: 4

Posted: Mon Jul 17, 2006 1:09 pm Post subject:
symmetry in eigenvectors



Hi!
Maybe someone could help me with the following. It would be very
helpful!
Given a real symmetric 4x4 matrix A (it's a covariance matrix), and one
real eigenvector x=(x1,x2,x3,x4).
Is it now possible to show that y=(x2,x1,x4,x3) is an eigenvector of
A, too?
Thanks,
Al 

Back to top 


Google


Back to top 



The time now is Wed Jul 18, 2018 10:18 pm  All times are GMT

