11-Point Interpolated AP

Problem

Compute the 11-point interpolated average precision (VOC 2007 style) from score-ordered labels.

For recall levels 0.0, 0.1, ..., 1.0 take the maximum precision achieved at a recall of at least that level; labels[k] is 1 for a positive and 0 otherwise, already sorted by descending score. Return the mean of the 11 values rounded to 4 decimal places, or 0.0 when there are no positives.

Test cases

case 1

in: [[1,0,1,1,0]]

expected: 0.8409

case 2

in: [[0,0,0]]

expected: 0

case 3

in: [[1,1,0,0,1]]

expected: 0.8545

case 4

in: [[1]]

expected: 1

Starter codePython

def ap_11_point(labels):
    # Your code here
    pass

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