Weight of Evidence

Problem

Compute the weight of evidence for one bin: ln((good_c/good_total) / (bad_c/bad_total)), clipping each distribution share to at least 1e-6.

Return 0.0 if a total is 0.

Hint

WOE compares the good and bad distributions inside a bin.

case 1

in: [30,10,100,100]

expected: 1.0986122886681096

case 2

in: [20,40,100,100]

expected: -0.6931471805599453

case 3

in: [50,50,100,100]

expected: 0

case 4

in: [0,10,100,100]

expected: -11.512925464970229

Starter codePython

import math
def woe(good_c, bad_c, good_total, bad_total):
    # Your code here
    pass

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