Cost-Sensitive Decision Threshold

Problem

Compute the cost-sensitive decision threshold for predicting the positive class: cost_fp / (cost_fp + cost_fn). Predict positive when the estimated probability exceeds this break-even point. Return 0.0 when the two costs sum to 0 or less.

Hint

When false negatives are expensive the threshold drops below 0.5 to catch more positives.

Test cases

case 1

in: [1,4]

expected: 0.2

case 2

in: [5,5]

expected: 0.5

case 3

in: [0,3]

expected: 0

case 4

in: [0,0]

expected: 0

case 5

in: [3,1]

expected: 0.75

Starter codePython

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

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