Repeated CV Variance

Problem

Compute the between-repeat variance of repeated cross-validation. Average the fold scores inside each repeat, then return the population variance of those repeat means. Fewer than two non-empty repeats returns 0.0.

Hint

High variance across repeats means the reported score depends heavily on the random split.

Test cases

case 1

in: [[[0.8,0.9],[0.7,0.8],[0.75,0.85]]]

expected: 0.0016666666666666698

case 2

in: [[[1],[0],[0.5]]]

expected: 0.16666666666666666

case 3

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

expected: 0

case 4

in: [[[]]]

expected: 0

Starter codePython

def repeated_cv_variance(repeat_scores):
    # Your code here
    pass

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