Position Bias Correction

Problem

Correct click-through rate for position bias with inverse propensity weighting: the mean of clicks_i / max(examination_prob_i, 1e-6). Empty input returns 0.0.

Hint

Dividing by the probability that a position was examined removes its advantage.

Test cases

case 1

in: [[1,0,1],[1,0.5,0.25]]

expected: 1.6666666666666667

case 2

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

expected: 1

case 3

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

expected: 0

case 4

in: [[],[]]

expected: 0

Starter codePython

def position_bias_corrected_ctr(clicks, examination_probs):
    # Your code here
    pass

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