MediumNLPnlp-179

Confusion Pair Mining

Problem

Find the most frequent ordered confusion pair. Return [gold, predicted, count] with the highest count, breaking ties by gold and then predicted alphabetically.

Return an empty list when there are no errors.

Hint

Iterate the sorted pair keys and keep the first strict maximum.

Test cases

case 1

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

expected: ["a","b",1]

case 2

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

expected: ["a","b",2]

case 3

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

expected: []

case 4

in: [[],[]]

expected: []

Starter codePython

def confusion_pair_mining(gold_labels, predicted_labels):
    # Your code here
    pass

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