How to get N maximum values in a multi dimensional numpy array along a given axis(say 2)?

Question

Since argmax only gives one maximum values,how can we find atleast 2 or 3 elements instead of just one.

Currently my input is in the format np.argmax(array,axis=2) which is giving only one maximum and i have to extract 2 or 3 atleast from the array which is N-dimensional

What do you mean by "2 or 3 elements"? np.argmax extracts the argmax of all arrays in one dimension. Are you looking for the highest, second highest and third highest? — Martino, Jun 17 '18 at 10:57
Question has nothing to do with `machine-learning` - kindly do not spam the tag (removed) — desertnaut, Jun 17 '18 at 11:08
Possible duplicate of https://stackoverflow.com/questions/6910641/how-to-get-indices-of-n-maximum-values-in-a-numpy-array — Merlin1896, Jun 17 '18 at 11:25

score 1 · Answer 1 · answered Jun 17 '18 at 12:10

I would try to use the function called argpartition(). To get the indices of the two largest elements, do:

import numpy as np

a = np.array([9, 4, 4, 3, 3, 9, 0, 4, 6, 0])

ind = np.argpartition(a, -2)[-2:] 

ind
Out[13]: array([5, 0], dtype=int64)

a[ind]
Out[14]: array([9, 9])

score 1 · Answer 2 · answered Jun 17 '18 at 12:21

1

Using numpy.argsort. Data from @CarlesSansFuentes.

import numpy as np

a = np.array([9, 4, 4, 3, 3, 9, 0, 4, 6, 0])

args = np.argsort(-a)[:2]

array([0, 5], dtype=int64)

answered Jun 17 '18 at 12:21

jpp

159,742
34
281
339

How to get N maximum values in a multi dimensional numpy array along a given axis(say 2)?

2 Answers2