[SciPy-user] determinant of a sparse matrix

Andy Fraser afraser@lanl....
Tue Dec 23 11:14:02 CST 2008


>>>>> "G" == Giampaolo Cuoghi <jpaul74@gmail.com> writes:

    G> Andy Fraser <afraser <at> lanl.gov> writes:
    >> 
    >> >>>>> "GC" == Giampaolo Cuoghi <jpaul74 <at> gmail.com> writes:
    >> 
    GC> In a particular problem I haven't to solve the linear system
    GC> Ax=b (with A=sparse matrix) but only compute the determinant
    GC> of A, and compare it with zero.
    >> 
    GC> Is it possible, in scipy, the computation of the determinant
    GC> of a sparse matrix?
    >> 
    >> I use scipy.sparse.linalg.lobpcg to find the smallest
    >> eigenvalue of a sparse matrix.  If you only want to know if
    >> your determinant is different from zero, and you know something
    >> about the eigenvalues (for example they are real and not
    >> negative) you might be able to get what you want out of lobpcg.
    >> 
    >> Andy
    >> 

Hi Andy,

    G> lobpcg is implemented only in scipy 0.7? It works only for
    G> symmetric real and hermitean complex matrix?

    G> I'm using scipy 0.6 and my matrix is an unsimmetryc complex
    G> sparse one.

Then my suggestion will not help.  For my application, I want to know
if a proposed covariance function (real and symmetric) is positive
definite and if not, how close it is.  I use lobpcg which I doscovered
at http://docs.scipy.org/doc/scipy/reference/sparse.linalg.html Since
I use intrepid ubuntu, I had to build numpy and scipy from svn.

Andy


More information about the SciPy-user mailing list