HardNLPnlp-358

Weighted Jaccard Token Similarity

Problem

The weighted Jaccard (Ruzicka) similarity of two token multisets is sum_t min(count_a, count_b) / sum_t max(count_a, count_b); define it as 1.0 when both multisets are empty.

Given two token lists, return the similarity.

Hint

Weight by counts and compare min against max per token.

Test cases

case 1

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

expected: 0.5

case 2

in: [["x"],["y"]]

expected: 0

case 3

in: [[],[]]

expected: 1

Starter codePython

def weighted_jaccard_token_similarity(tokens_a, tokens_b):
    # Your code here
    pass

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