EasyNLPnlp-285

POS Tagging Accuracy with Ignore Set

Problem

Compute POS tagging accuracy while ignoring selected gold tags. gold_tags and pred_tags are parallel label lists, and ignore is the set of gold tags to skip.

The lists are zipped to the shorter length. Return correct non-ignored predictions divided by the number of non-ignored positions, or 0.0 when none are counted.

Hint

Skip positions whose gold tag is in the ignore set before counting.

Test cases

case 1

in: [["N","V","P"],["N","N","P"],[]]

expected: 0.6666666666666666

case 2

in: [["N","V","P"],["N","N","P"],["P"]]

expected: 0.5

case 3

in: [["N"],[],[]]

expected: 0

case 4

in: [[],[],[]]

expected: 0

case 5

in: [["P","P"],["N","N"],["P"]]

expected: 0

Starter codePython

def pos_tagging_accuracy(gold_tags, pred_tags, ignore):
    # Your code here
    pass

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