Faster way of calculating a distance matrix with numpy?

Question

I am calculating a matrix with numpy/scipy like this:

cost = np.empty([chroma1.data.shape[1], chroma2.data.shape[1]])

for x, cx in enumerate(chroma1.transpose()):
    for y, cy in enumerate(chroma2.transpose()):
        cost[x, y] = sp.distance.euclidean(cx, cy)

This takes quite an amount of time. Is there any numpy/scipy function that would allow me to get rid of the two nested loops?

Joe Kington · Accepted Answer · 2016-01-27 16:00:19Z

3

It looks like you're calculating a distance matrix. scipy.spatial.distance contains several specialized, optimized functions for doing exactly that.

In your case:

cost = scipy.spatial.distance.cdist(chroma1.T, chroma2.T)

should do exactly what you want.

answered Jan 27, 2016 at 16:00

Joe Kington

287k73 gold badges621 silver badges474 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Faster way of calculating a distance matrix with numpy?

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related