Cost-Weighted Cutoff

Problem

Return the probability cutoff c_fp / (c_fp + c_fn) that minimizes expected cost when a false positive costs cost_fp and a false negative costs cost_fn. Return 0.0 when both costs are zero.

Hint

Predict positive when the expected cost of doing so is lower.

Test cases

case 1

in: [1,1]

expected: 0.5

case 2

in: [4,1]

expected: 0.8

case 3

in: [1,4]

expected: 0.2

case 4

in: [0,0]

expected: 0

case 5

in: [2,8]

expected: 0.2

Starter codePython

def cost_weighted_cutoff(cost_fp, cost_fn):
    # Your code here
    pass

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