Uplift Calibration Error

Problem

Compute the count-weighted calibration error of uplift predictions: sum over bins of (count/total) * |predicted - observed|.

Return 0.0 when the total count is 0.

Hint

Calibration checks whether predicted uplift magnitude matches reality.

case 1

in: [[0.1,0.2],[0.05,0.3],[100,50]]

expected: 0.06666666666666665

case 2

in: [[0.2],[0.2],[10]]

expected: 0

case 3

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

expected: 0

Starter codePython

def uplift_calibration_error(predicted_uplift, observed_uplift, counts):
    # Your code here
    pass

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