MediumNLPnlp-028

Add-k Smoothing

Problem

Return the add-k smoothed conditional probability (count + k) / (context_count + k * vocab_size).

count is the observed count of the event, context_count is the total count of its context, vocab_size is the number of possible outcomes, and k is the smoothing constant.

Hint

Add k to the numerator and k times the vocabulary size to the denominator.

Test cases

case 1

in: [0,10,5,1]

expected: 0.06666666666666667

case 2

in: [3,10,5,1]

expected: 0.26666666666666666

case 3

in: [0,0,4,0.5]

expected: 0.25

case 4

in: [5,5,1,0.1]

expected: 1

Starter codePython

def add_k_smoothing(count, context_count, vocab_size, k):
    # Your code here
    pass

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