EasyNLPnlp-245

BM25 IDF Computation

Problem

Return the BM25 probabilistic inverse document frequency log(1 + (n_docs - df + 0.5) / (df + 0.5)).

Return 0.0 when n_docs is not positive or df is negative or greater than n_docs.

Hint

This idf is always positive, unlike the classic log(n_docs / df).

case 1

in: [10,100]

expected: 2.2637452596777816

case 2

in: [1,1]

expected: 0.28768207245178085

case 3

in: [0,10]

expected: 3.091042453358316

case 4

in: [5,5]

expected: 0.0870113769896297

case 5

in: [11,10]

expected: 0

Starter codePython

import math
def bm25_idf(df, n_docs):
    # Your code here
    pass

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