Re: ? determine SVD of very large matrix


"Carl Barron" <cbarron413@xxxxxxxxxxxx> wrote in message
news:130420060629413069%cbarron413@xxxxxxxxxxxxxxx
In article <dIm%f.135$cs4.18@xxxxxxxxxxxxxxxx>, Cheng Cosine
<acosine@xxxxxxxxxxxx> wrote:
...
I have a pdf entitled "Out-of-Core SVD and QR Decompositions" by
Eran Rabani and Sivan Toledo, I got from the internet.

found somewhere with a google svd. but using this information you
should be able to shorten the search.


Yes, their approach somehow share a basic idea like Toni Lassila mentioned:

try to work on a smaller matrix.

But how to deal with large full-randked square matrix? In solving large
linear system,

one can find some iterative methods to avoid using large computer memory,
are there

iterative method for solving SVD?

Thanks,
by Cheng Cosine
Apr/14/2k6 NC


.

Quantcast