Subject: Computing SVD Of "Large Sparse Data"

I think that incorporating mean subtraction into the SSVD code should
be relatively straightforward.  The trick is that you have to project
the orginal matrix and the mean separately and then combine the
results.  There are later steps that require similar mods, but since
it is all based on a simple product that can always be factored, it
should be straightforward.

On Tue, Jun 7, 2011 at 4:06 PM, Eshwaran Vijaya Kumar