Numpy - referencing values in other array for multidimensional array

Question

I'm writing a k nearest neighbor classifier in Numpy and faiss (Facebook kNN library). For point classification, I receive:

[[ 9240  4189  8702]
 [ 2639  1052 13565]
 [10464 14220 13980]
 ...
 [12014 12063  1331]
 [ 6719  5832  8827]
 [ 1793  5455 12328]]

Each row is an index of the y vector. I need to reference the values for this matrix in the y vector, so e. g. I will swap 9240 in the matrix for y[9240] value, for example 1 (positive class).

Can I do that without Python loop, i. e. can this be done with Numpy only?

@alani Yeah, that's it! For some reason I thought it would not work for multicolumn X... please post an answer so I can accept. — qalis
– qalis, Commented Sep 7, 2020 at 13:09

alani · Accepted Answer · 2020-09-07 13:11:51Z

1

You can index a 1-d array with an array of indices. The result is the same shape as the array of indices. For example:

>>> a
array([[1, 2, 2],
       [3, 3, 2]])

>>> y
array([ 1000.,  1020.,  1040.,  1060.])

>>> y[a]
array([[ 1020.,  1040.,  1040.],
       [ 1060.,  1060.,  1040.]])

answered Sep 7, 2020 at 13:11

alani

13.2k3 gold badges18 silver badges34 bronze badges

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

Collectives™ on Stack Overflow

Numpy - referencing values in other array for multidimensional array

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