sparse matrix division (least squares)

Rishi Amrit rishiamrit at gmail.com
Tue Mar 31 14:02:24 CDT 2009


On Tue, Mar 31, 2009 at 12:14 AM, David Bateman <dbateman at dbateman.org>wrote:

> Rishi Amrit wrote:
>
>  On Mon, Mar 30, 2009 at 6:02 PM, Ben Abbott <bpabbott at mac.com <mailto:
>> bpabbott at mac.com>> wrote:
>>
>>
>>    On Mar 30, 2009, at 12:05 PM, Rishi Amrit wrote:
>>
>>        Hi,
>>
>>          I have two sparse matrices (say A and B).
>>
>>        > size(A)
>>        ans =
>>          2400  2400
>>
>>        > size(B)
>>        ans =
>>          2400 2440
>>
>>        I wish to implement the least squares operator : A\B
>>
>>        When I run this in MATLAB, I get
>>
>>        > size(A\B)
>>        ans =
>>         2440  2400
>>
>>        and A\B is a sparse matrix.
>>
>>        When I implement this in octave, I get
>>
>>        > A\B
>>        ans = Compressed Column Sparse (rows = 0, cols =0, nnz = 0)
>>
>>        which is empty.
>>
>>        Is there a separate function in octave for this operation on
>>        sparse matrices ?
>>
>>        Thanks,
>>
>>        Rishi
>>
>>
>>    What version of Octave are you running?
>>
>>
>> I am running the latest stable version: 3.0.3
>>
>> Thanks,
>>
>> Rishi
>>
>>  Where did you get your version of Octave? Is it built with SuiteSparse? I
> don't see this issue with 3.0.3 on my machine
>


Installing libsuitesparse-dev and recompiling octave fixed the issue.

Thanks,

Rishi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://www-old.cae.wisc.edu/pipermail/help-octave/attachments/20090331/af6fe70e/attachment.html 


More information about the Help-octave mailing list