[SciPy-user] Re: [SciPy-dev] matrixmultiply()

Robert Kern rkern at ucsd.edu
Fri Apr 8 18:32:16 CDT 2005


David Grant wrote:
> Can anyone tell me why the * operation on matrices is for element-wise
> multiplcation and not matrix multiplication for Numeric/numpy arrays?

Not everyone is using Numeric for linear algebra on matrices. 
Element-wise multiplication is the most consistent interpretation for * 
in this general-purpose context.

Sine you *are* doing linear algebra on matrices, you can use Matrix 
objects, which do define the * operator to be matrix multiplication.

In [8]:A = mat([[0,1],[2,3]])

In [9]:B = mat([[5,6],[7,8]])

In [10]:A*B
Out[10]:
Matrix([[ 7,  8],
        [31, 36]])

In [11]:dot(A, B)
Out[11]:
array([[ 7,  8],
        [31, 36]])


> Or at least tell me a nice way to do A*B*C*D*E*F, where the * denotes
> matrix multiplication. I once wrote a function where I passed those to a
> function as arrary of arrays, so mult([A,B,C,D,E,F]) and the function
> multiplied them all together. But I can't remember where I put this
> function, and I was wondering if there was an easier way.

Alternatively, you can do a bit of magic with this recipe:

http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/384122

In [25]:d = Infix(dot)

In [26]:a = array([[0,1],[2,3]])

In [27]:b = array([[5,6],[7,8]])

In [28]:dot(a,b)
Out[28]:
array([[ 7,  8],
        [31, 36]])

In [29]:a |d| b
Out[29]:
array([[ 7,  8],
        [31, 36]])

-- 
Robert Kern
rkern at ucsd.edu

"In the fields of hell where the grass grows high
  Are the graves of dreams allowed to die."
   -- Richard Harter



More information about the SciPy-user mailing list