Get top N values from each sub-array from 2D numpy array

Asked Jun 04 '18 at 09:04

Active Jun 04 '18 at 09:04

Viewed 60 times

I've a 2 dimensional numpy array (Classification Model's probability output). Now I want to extract the index of top 3 values for each nested array.

E.g.

x = np.array([[1, 2, 3, 5, 6, 8], [4, 5, 6, 9, 10, 1]], np.int32)

What I want is

[[5, 4, 3]
[4, 3, 2]]

Without using loop

asked Jun 04 '18 at 09:04

Aditya Thakkar

@asapokL Not really, OP is asking about how to retrieve the position of maximum numbers without loop, actually a method problem, related to unravel_index method. – MT-FreeHK Jun 04 '18 at 09:10
1

np.argsort or argposition applied to the axis? – hpaulj Jun 04 '18 at 09:15
E.g. `np.argsort(x, axis=1)[:, :-4:-1]` – jdehesa Jun 04 '18 at 09:56
@jdehesa thanks your solution worked!!! – Aditya Thakkar Jun 04 '18 at 10:18
@asapokL please understand the question properly. There is no easy language than this, I could have explained. – Aditya Thakkar Jun 04 '18 at 10:20

0 Answers0