[SciPy-User] Sparse matrix multiply
Jaakko Luttinen
Why do I get two different results for the code below?
import numpy as np
import scipy.sparse as sp
A = sp.rand(20,20,density=0.1)
B = sp.rand(20,20,density=0.1)
np.multiply(A,B).sum()
# out: 21.058793740984925
A.multiply(B).sum()
# out: 0.76482546226069481
Am I missing something?
I think numpy.multiply should either return the correct answer or an
error that it can't compute the correct answer.
Regards,
Jaakko
_______________________________________________
SciPy-User mailing list
SciPy...@scipy.org
http://mail.scipy.org/mailman/listinfo/scipy-user
Christopher Mutel
<jaakko....@aalto.fi> wrote:
> Hi!
>
> Why do I get two different results for the code below?
>
> import numpy as np
> import scipy.sparse as sp
> A = sp.rand(20,20,density=0.1)
> B = sp.rand(20,20,density=0.1)
> np.multiply(A,B).sum()
> # out: 21.058793740984925
> A.multiply(B).sum()
> # out: 0.76482546226069481
>
> Am I missing something?
> I think numpy.multiply should either return the correct answer or an
> error that it can't compute the correct answer.
np.multiply performs element-wise multiplication, while A.multiply is
matrix multiplication. They are both "correct", but answer different
questions.
See:
http://en.wikipedia.org/wiki/Matrix_multiplication
http://docs.scipy.org/doc/numpy/reference/generated/numpy.multiply.html
> Regards,
> Jaakko
> _______________________________________________
> SciPy-User mailing list
> SciPy...@scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-user
--
############################
Chris Mutel
Ökologisches Systemdesign - Ecological Systems Design
Institut f.Umweltingenieurwissenschaften - Institute for Environmental
Engineering
ETH Zürich - HIF C 44 - Schafmattstr. 6
8093 Zürich
Telefon: +41 44 633 71 45 - Fax: +41 44 633 10 61
############################
Jaakko Luttinen
> On Thu, Mar 22, 2012 at 10:34 AM, Jaakko Luttinen
> <jaakko....@aalto.fi> wrote:
>> Hi!
>>
>> Why do I get two different results for the code below?
>>
>> import numpy as np
>> import scipy.sparse as sp
>> A = sp.rand(20,20,density=0.1)
>> B = sp.rand(20,20,density=0.1)
>> np.multiply(A,B).sum()
>> # out: 21.058793740984925
>> A.multiply(B).sum()
>> # out: 0.76482546226069481
>>
>> Am I missing something?
>> I think numpy.multiply should either return the correct answer or an
>> error that it can't compute the correct answer.
>
> np.multiply performs element-wise multiplication, while A.multiply is
> matrix multiplication. They are both "correct", but answer different
> questions.
Is it so..?
http://docs.scipy.org/doc/scipy/reference/generated/scipy.sparse.csc_matrix.multiply.html
I don't know what "point-wise multiplication" means..
Anyway, I thought that dot computes matrix multiplication and multiply
computes matrix multiplication.
-Jaakko
Jaakko Luttinen
TYPOFIX: I thought that multiply computes element-wise multiplication.
Pauli Virtanen
> import numpy as np
> import scipy.sparse as sp
> A = sp.rand(20,20,density=0.1)
> B = sp.rand(20,20,density=0.1)
> np.multiply(A,B).sum()
> Am I missing something?
> I think numpy.multiply should either return the correct answer or an
> error that it can't compute the correct answer.
The answer is the same as to your previous questions --- the Numpy
ufuncs do not deal with sparse matrices in a reasonable way. This lack
of integration between dense and sparse is a bug.
Why it does not raise an error instead, is probably that as a
consequence of the operation overloading rules defined, there is a
(nonsensical) operation that matches the call.
--
Pauli Virtanen
Jaakko Luttinen
> <jaakko....@aalto.fi> wrote:
>> Hi!
>>
>> Why do I get two different results for the code below?
>>
>> import numpy as np
>> import scipy.sparse as sp
>> A = sp.rand(20,20,density=0.1)
>> B = sp.rand(20,20,density=0.1)
>> np.multiply(A,B).sum()
>> # out: 21.058793740984925
>> A.multiply(B).sum()
>> # out: 0.76482546226069481
>>
>> Am I missing something?
>> I think numpy.multiply should either return the correct answer or an
>> error that it can't compute the correct answer.
>
> np.multiply performs element-wise multiplication, while A.multiply is
> matrix multiplication. They are both "correct", but answer different
> questions.
Actually it seems that np.multiply computes matrix multiplication in
this case! Below, only A.multiply(B) computes element-wise multiplication.
import numpy as np
import scipy.sparse as sp
A = sp.rand(20,20,density=0.1)
B = sp.rand(20,20,density=0.1)
np.multiply(A,B).sum()
# out: 25.240683127057885
A.multiply(B).sum()
# out: 2.6382118196920503
A.dot(B).sum()
# out: 25.240683127057885
np.dot(A,B).sum()
# out: 25.240683127057885
-Jaakko
Christopher Mutel
Indeed. Sorry for the confusion.