Chi-Square Drift Test

Problem

Compute the chi-square drift statistic between expected and actual counts: sum over bins with expected > 0 of (actual - expected)^2 / expected.

Empty input returns 0.0.

Hint

Large statistics indicate the actual distribution has drifted from expected.

case 1

in: [[50,30,20],[40,40,20]]

expected: 5.333333333333334

case 2

in: [[50,30,20],[50,30,20]]

expected: 0

case 3

in: [[0,10],[5,5]]

expected: 2.5

case 4

in: [[],[]]

expected: 0

Starter codePython

def chi_square_drift(expected_counts, actual_counts):
    # Your code here
    pass

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