[Scilab-users] Generalised eigenvalues

[Serge Steer]

Le 21/12/2012 10:53, Dang, Christophe a écrit :
> Hello,
>
> I'd like to convert the following Matlab script
>
> [gevec, geval] = eig(S, C);
>
> (see http://homepages.inf.ed.ac.uk/rbf/CVonline/LOCAL_COPIES/FITZGIBBON/ELLIPSE/ )
>
> The MathWorks page
> http://www.mathworks.fr/fr/help/matlab/ref/eig.html
> tells me these are the generalised eigen-values and-vectors.
>
> According to the Scilab help page, spec seems to works a bit differently.
>
> I guess I have to do something like
>
> [al,be,gevec] = spec(S,C)
> geval = al./be
>
> but this is a bit beyond my math skills.
>
> Am I right or is the truth "out there"?
You are right.  but Scilab returns the generalized eigenvalues under a 
factorized form because some generalized eigenvalue problems may have 
eigenvalue at infinity.

Serge Steer