Hola!
This can be done with the CRAN package igraph, which contains (part
of) the arpack
library for computing only some eigenvalues/eigenvectors of sparse
matrices. arpack gives you the option of computing a few of the
smallest or a few of the largest eigenvalues/vectors.
You will need yourself to wrte a function doing matrix-vector
multiplication, so the arpack
methods itself is independent of the implementation of your sparse matrix.
Kjetil
On Fri, Mar 9, 2012 at 9:09 AM, Holger Diedrich
<diedrich at math.uni-potsdam.de> wrote:> Dear all,
>
> I am currently working on the calculation of eigenvalues (and -vectors) of
> large matrices. Since these are mostly sparse matrices and I remember some
> specific functionalities in MATLAB for sparse matrices, I started a
research
> how to optimize the calculation of eigenvalues of a sparse matrix.
> The function eigen itself works with the LAPACK library which has no
special
> handling for sparse matrices, same for the EISPACK library. The ARPACK
> library is capable to work with sparse matrices but I couldn't find any
> useful R package using ARPACK.
> The Matrix package can handle sparse matrices but has no further useful
> functionalities (concerning my tasks).
>
> Does one of you have any advice how to optimize the eigenvalue calculation
> of sparse matrices in R?
>
> Thanks in advance,
> Holger
>
> ______________________________________________
> R-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel