PR AUC Trapezoid

Problem

Compute the area under a precision-recall curve with the trapezoidal rule. For each distinct recall keep the maximum precision, sort by recall, and sum (r1 - r0) * (p0 + p1) / 2.

Return 0.0 with fewer than two distinct recall values.

Hint

Keeping one precision per recall first removes vertical segments from the curve.

Test cases

case 1

in: [[[1,0.5],[0.5,0.5],[0.6666666666666666,1],[0.5,1]]]

expected: 0.41666666666666663

case 2

in: [[[1,1]]]

expected: 0

case 3

in: [[[0.9,0.2],[0.7,0.5],[0.5,1]]]

expected: 0.54

case 4

in: [[]]

expected: 0

Starter codePython

def pr_auc_trapezoid(pr_points):
    # Your code here
    pass

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