Le 17/05/2011 20:40, Arne Plöse a écrit :
Is there a Eigenvalue decomposition like octave "eig(m)" that takes also
matices that are not symmetric? Or any mathematical tip how to do this?
No, we do not have an implementation for non-symmetric matrices.
There is an open issue for that
<https://issues.apache.org/jira/browse/MATH-235>. We would be happy to
have one (and we would be happy to fix our SVD implementation too).
Algorithms to do this generally invold first a transform to upper
Hessenberg (i.e. upper triangular with additional non-null terms just
below the main diagonal) and then a Schur decomposition. I think Schur
decomposition implies working with complex elements.
Patch welcome!
Luc
Arne
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
