Vectorization of Euclidean distance calculation

Arvind Iyer — Sat, 08 Nov 2008 23:10:04 -0500

The norm command provides a quick way of calculating Euclidean distance in an N-dimensional space.
e.g,
N = 50;
x1 = randn(1,N);
x2 = randn(1,N);
euclid_dist = norm(x1-x2);

Consider I have a M points x1, x2.....xM stored in an MxN matrix
[x1;x2.....;xM]
I now want to find the Euclidean distance of each of these point from a test point xT.

How can this be done in a vectorized way without loops?
I am interested in speeding this up because M is really large in my problem.

Re: Vectorization of Euclidean distance calculation

John D'Errico — Sat, 08 Nov 2008 23:58:02 -0500

"Arvind Iyer" <aiyer@ict.usc.edu> wrote in message <gf568c$ft5$1@fred.mathworks.com>...
> The norm command provides a quick way of calculating Euclidean distance in an N-dimensional space.
> e.g,
> N = 50;
> x1 = randn(1,N);
> x2 = randn(1,N);
> euclid_dist = norm(x1-x2);
>
> Consider I have a M points x1, x2.....xM stored in an MxN matrix
> [x1;x2.....;xM]
> I now want to find the Euclidean distance of each of these point from a test point xT.
>
> How can this be done in a vectorized way without loops?
> I am interested in speeding this up because M is really large in my problem.

http://www.mathworks.com/matlabcentral/fileexchange/18937

John

MATLAB Central Newsreader - Vectorization of Euclidean distance calculation

Vectorization of Euclidean distance calculation

Re: Vectorization of Euclidean distance calculation