Permutation Importance Rank

Problem

Rank features by mean permutation importance. Average each column of the matrix across repeats, then return feature indices ordered by descending mean, breaking ties by the smaller index. Empty input returns [].

Hint

Averaging over repeats makes the ranking less sensitive to a single unlucky shuffle.

Test cases

case 1

in: [[[0.1,0.4,0.2],[0.2,0.5,0.3]]]

expected: [1,2,0]

case 2

in: [[[0.5,0.5,0.1]]]

expected: [0,1,2]

case 3

in: [[[1,2],[2,1]]]

expected: [0,1]

case 4

in: [[]]

expected: []

case 5

in: [[[],[]]]

expected: []

Starter codePython

def permutation_importance_rank(importance_matrix):
    # Your code here
    pass

5 test cases · run your code against them in the editor