Feature Bagging Indices

Problem

Split feature indices into random subsets. Call random.seed(seed), shuffle range(n_features), and cut it into consecutive chunks of size k.

Return each chunk sorted; a non-positive k returns [].

Hint

Random feature subsets decorrelate the trees in a forest.

case 1

in: [6,3,42]

expected: [[1,2,3],[0,4,5]]

case 2

in: [5,2,7]

expected: [[0,4],[1,3],[2]]

case 3

in: [4,4,0]

expected: [[0,1,2,3]]

case 4

in: [3,1,1]

expected: [[1],[2],[0]]

Starter codePython

import random
def feature_bagging_indices(n_features, k, seed):
    # Your code here
    pass

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