Cohen's Kappa Agreement

Problem

Return Cohen's kappa (po - pe) / (1 - pe) for two raters, where po is observed agreement and pe is the sum over labels of the product of each rater's label marginals. Return 0.0 when the inputs are empty, have different lengths, or pe equals 1.0.

Test cases

case 1

in: [["a","b","a","b"],["a","b","b","b"]]

expected: 0.5

case 2

in: [["a","a"],["a","a"]]

expected: 0

case 3

in: [["a","b"],["b","a"]]

expected: -1

case 4

in: [[],[]]

expected: 0

Starter codePython

def cohens_kappa(a, b):
    # Your code here
    pass

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