8
$\begingroup$

Given two vectors of size $n$ $$u = [u_1, u_2, u_3, ..., u_n ] $$ and $$v = [v_1, v_2, v_3, ..., v_n ] $$

What is the name of the operation "$u ? v$" such that the result is a vector of size $n$ of the form $$u ? v = [v_1 \times u_1, v_2 \times u_2, v_3\times u_3, ..., v_n \times u_n ]$$

For want of a better name, I have termed it "piecewise vector multiplication".

What is this operation normally known as in the literature?

Improve this question
$\endgroup$
1
  • 8
    $\begingroup$ "Componentwise" or "coordinatewise" sounds more like standard terminology than "piecewise", but I don't know a really common name for this. $\endgroup$
    – darij grinberg
    Commented Dec 17, 2009 at 11:40

4 Answers 4

Reset to default
8
$\begingroup$

It's pointwise product. See Wikipedia articles here and here

Improve this answer
$\endgroup$
1
  • $\begingroup$ I have chosen this as the correct answer since the links provided better information than what is produced by a search for the (more) correct answer "Hadamard product". For example, the Mathworld article on the Hadamard product is just confusing in this context. $\endgroup$
    – Niall Murphy
    Commented Dec 17, 2009 at 17:55
22
$\begingroup$

That is the Hadamard product---which usually, though, is only used with matrices of a more matrixy shape.

Improve this answer
$\endgroup$
2
  • 7
    $\begingroup$ I heard a seminar by a functional analyst where this product figured quite prominently, and throughout which it was referred to systematically as «the wrong product»... $\endgroup$
    – Mariano Suárez-Álvarez
    Commented Dec 17, 2009 at 12:00
  • 7
    $\begingroup$ "The freshman product". :-) $\endgroup$
    – Greg Kuperberg
    Commented Dec 17, 2009 at 14:35
13
$\begingroup$

If $k$ is a field, the vector space $k^n$ endowed with the componentwise multiplication is called a diagonal algebra ( and so are isomorphic algebras). The terminology is due to Bourbaki and is justified by the following result.

If $M$ is a square matrix over $k$, it is diagonalizable over $k$ if and only if the algebra $k[M]$ is diagonal.The proof results from the diagonalization criterion (the minimal polynomial of $M$ should be split over $k$ and have distinct roots) , the isomorphism of $k$-algebras $ \frac {k[X]} {polmin_M (X)}\to k[M]$ and the Chinese remainder theorem.

These algebras are important to algebraic geometers because they are a model for étale algebras over a field. Indeed, a $k$ -algebra $A$ is étale if and only it becomes diagonal after some extension of the base field . More explicitly $A$ is étale if and only if for some field extension $K/k$ we have an isomorphism $A\otimes_k K \simeq K^n$ of K-algebras.

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ +1 mainly for the Bourbaki mention (also because this is correct.) $\endgroup$
    – Harry Gindi
    Commented Dec 17, 2009 at 15:07
2
$\begingroup$

If you look at u and v as functions on the set S={1,2,3,...,n}, then they are elements of a function space on S, which is an algebra under pointwise addition and multiplication.

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .