Decision Curve Net Benefit

Problem

Return the net benefit tp/n - (fp/n) * (threshold / (1 - threshold)) for a decision threshold. Return 0.0 when n <= 0 or the threshold is not strictly between 0 and 1.

Hint

It compares the value of true positives against the harm of false positives at the given threshold.

Test cases

case 1

in: [30,10,100,0.2]

expected: 0.27499999999999997

case 2

in: [50,50,100,0.5]

expected: 0

case 3

in: [0,0,100,0.3]

expected: 0

case 4

in: [100,0,100,0.1]

expected: 1

case 5

in: [10,10,100,0]

expected: 0

Starter codePython

def decision_curve_net_benefit(tp, fp, n, threshold):
    # Your code here
    pass

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