Re: SVD(M.transpose())

[Cleve Moler <moler@mathworks.com>]

Hi.

You are trying to find the null space of a matrix with more columns
than rows.  Is that correct?  The version of SVD currently available
in JAMA can not do that.  The JAMA SVD doesn't work on matrices
with n > m, so I suggested you take the SVD of the transpose.
But the JAMA SVD only computes the "economy sized" SVD.  So, as
you observed, the U matrix doesn't have the n-m "extra" columms
that would give you the null space.  A more complete SVD function
would have some kind of option for producing either the economy sized
or the full SVD, and would work on matrices of all shapes.  We
haven't done that yet.

  -- Cleve Moler