Probability Smoothing

Definition

Probability smoothing is a language modeling technique that assigns some nonzero probability to events that were unseen in the training data. This has the effect that the probability mass is divided over more events; hence, the probability distribution becomes more smooth.

Key Points

Smoothing overcomes the so-called sparse data problem, that is, many events that are plausible in reality are not found in the data used to estimate probabilities. When using maximum likelihood estimates, unseen events are assigned a zero probability. In case of information retrieval, most events are unseen in the data, even if simple unigram language models are used documents that are relatively short (say on average several hundreds of words), whereas the vocabulary is typically big (maybe millions of words), so the vast majority of words does not occur in the document. A small document about “information retrieval” might not mention the word “search,” but that does not mean it is not relevant...