Page 1 of 1

Eigenvalue decomposition

Posted: Mon May 06, 2013 11:06 am
by mcorozcos
Hello Tom,

I have a question, I used the instruction “%eigdecomp” in order to calculate the eigenvalues and the eigenvectors for one matrix (because my matrix isn´t symmetric, for this reason I can´t use Eigen), but only I need are the eigenvalues, ¿How do I extract the eigenvalues and eigenvectors separated, using the instruction "%eigdecomp"?

I hope you can help me

Thanks

Re: Eigenvalue decomposition

Posted: Mon May 06, 2013 11:31 am
by TomDoan
You can (and in fact have to) use EIGEN for a non-symmetric matrix. You can use the CVALUES option if the eigenvalues might be complex, or just use the eigenvalues (2nd) parameter if they will be real.

Re: Eigenvalue decomposition

Posted: Mon May 06, 2013 1:25 pm
by mcorozcos
But, when I used the instruction "eigen" only I can use with square matrices.

In my case I have an other type of matrix, and When I use this instruction, the program shows error like "## MAT3. Matrix with Dimensions 486 x 488 Involved in EIGEN Operation. Need NxN" for this reason I think that I have to use the other instruction.

Or exist other instruction what can I use for calculate eigenvalues and eigenvectors a matrices no square?

Re: Eigenvalue decomposition

Posted: Mon May 06, 2013 1:35 pm
by TomDoan
No, because non-square matrices don't have eigenvalues. They have singular values, which can be computed with %SVDECOMP, but those don't really work the same way as eigenvalues.