Does this kind of matrix product exist?

In summary, the conversation discusses the existence of a product A * B, where A and B are matrices, that results in a matrix with elements equal to the product of the corresponding elements in A and B. This product is commonly used in MATLAB and is a special case of the tensor product. It can also be defined using index notation and is sometimes referred to as the "pointwise", "componentwise", "elementwise", or "Hadamard" product.
  • #1
lets_resonate
15
0
Suppose you have a matrix A:

[tex]
\left(
\begin{array}{ccc}
a_{1,1} & a_{1,2} & a_{1,3} \\
a_{2,1} & a_{2,2} & a_{2,3} \\
a_{3,1} & a_{3,2} & a_{3,3}
\end{array}
\right)
[/tex]

And a matrix B:

[tex]
\left(
\begin{array}{ccc}
b_{1,1} & b_{1,2} & b_{1,3} \\
b_{2,1} & b_{2,2} & b_{2,3} \\
b_{3,1} & b_{3,2} & b_{3,3}
\end{array}
\right)
[/tex]

I want a product [itex]A \star B[/itex] that would result in:

[tex]
\left(
\begin{array}{ccc}
a_{1,1} b_{1,1} & a_{1,2} b_{1,2} & a_{1,3} b_{1,3} \\
a_{2,1} b_{2,1} & a_{2,2} b_{2,2} & a_{2,3} b_{2,3} \\
a_{3,1} b_{3,1} & a_{3,2} b_{3,2} & a_{3,3} b_{3,3}
\end{array}
\right)
[/tex]

Does such a product exist? What would be the name of it?
 
Mathematics news on Phys.org
  • #2
It certainly exists. For one thing, you just defined it! But regardless of that fact, this is a common product in MATLAB: v1.*v2, where v1 and v2 are vectors or matrices of the same size. I'm not sure if it has a specific name, probably because it's a special case of a more general operation, the tensor product.

I can define a rank-4 tensor with components

[tex]C^{ik}_{~~jl} = A^{i}_{~j}B^{k}_{~l}[/tex]

I can then contract over the indices k and j to get the matrix you want:

[tex]M^{i}_{~l} = C^{ik}_{~~kl} = A^{i}_{~k}B^{k}_{~l}[/tex]

In case you're unfamilar with index notation, we're basically specifying the components of the matrix (or, in general, higher dimensional objects called tensors) using indices to label the rows and columns (and the other higher dimensional directions of the object). Note that I've used "Einstein Notation", meaning a summation is implied: any time an index appears twice on a product or term it is being summed over. The tensor C I defined is a new object defined in terms of the components of two other objects. The "Contraction" I did, summing over elements of the C tensor, results in a lower-dimensional object, which in this case is the object you defined.

See: http://en.wikipedia.org/wiki/Einstein_notation , http://en.wikipedia.org/wiki/Tensor_product , http://en.wikipedia.org/wiki/Tensor

Note that the stuff on tensors may be a little more than you want or need for your purposes, depending on what they are.
 
Last edited by a moderator:
  • #3
Mute said:
I can then contract over the indices k and j to get the matrix you want:

[tex]M^{i}_{~l} = C^{ik}_{~~kl} = A^{i}_{~k}B^{k}_{~l}[/tex]
No, that is the ordinary product of two matrices.
 
  • #4
lets_resonate said:
What would be the name of it?
My first guess is that someone using such a product would call it the "pointwise product", "componentwise product", or "elementwise product", if they bothered to name it at all.
 
  • #5
Hadamard product is the name I've seen used for it.
 
  • #6
I've seen it used in C* algebra related material, though I can't recall what it was used for.
 

1. What is a matrix product?

A matrix product is a mathematical operation that involves multiplying two matrices together. It results in a new matrix with dimensions determined by the dimensions of the original matrices.

2. How is a matrix product calculated?

To calculate a matrix product, the number of columns in the first matrix must be equal to the number of rows in the second matrix. Each element in the resulting matrix is calculated by multiplying the corresponding elements in the rows and columns of the original matrices and then adding them together.

3. What is the purpose of a matrix product?

Matrix products are used in many fields, including mathematics, physics, engineering, and computer science. They are used to represent linear transformations, solve systems of linear equations, and perform other mathematical operations.

4. Are all matrices compatible for multiplication?

No, not all matrices are compatible for multiplication. As mentioned earlier, the number of columns in the first matrix must be equal to the number of rows in the second matrix for a matrix product to exist. If this condition is not met, the matrices cannot be multiplied together.

5. Can a matrix product always be calculated?

No, a matrix product may not always be calculated. In some cases, the dimensions of the resulting matrix may not be well-defined or the matrices may not be compatible for multiplication. Additionally, even if a matrix product exists, it may not be computationally feasible to calculate due to the large number of operations involved.

Similar threads

Replies
1
Views
885
  • Linear and Abstract Algebra
Replies
3
Views
1K
Replies
2
Views
2K
  • Math Proof Training and Practice
Replies
10
Views
1K
  • Differential Equations
Replies
3
Views
2K
Replies
2
Views
135
  • Math Proof Training and Practice
Replies
8
Views
1K
Replies
5
Views
4K
  • Other Physics Topics
Replies
6
Views
1K
Replies
4
Views
744
Back
Top