Re: ? determine SVD of very large matrix

On Thu, 13 Apr 2006 07:02:33 GMT, "Cheng Cosine"
<acosine@xxxxxxxxxxxx> wrote:

Suppose one has a very large matrix, well, large enough for computer at
hands

cannot store it, how does one determine its SVD? Will symm or sparsity make
the problem tractable?

Can you implement the action of the adjoint of A to an arbitrary
vector? If so, you can use Krylov-subspace methods for finding the
largest eigenvalues of AA*.
.

Relevant Pages

  • ? determine SVD of very large matrix
    ... cannot store it, how does one determine its SVD? ... Will symm or sparsity make ...
    (sci.math.num-analysis)
  • Re: SVD for a large sparse matrix
    ... So how will you expect to store a full version ... I don't have a small matrix whose SVD i'm trying to ... It will require 520 gigabytes to store ... You don't have that much memory, ...
    (comp.soft-sys.matlab)
Loading