Natural Language Processing · M2 · read-chapter

N-grams and language models

Unit 3, pages 62–108; essential probability revision

1. Words, counts and probabilities: the notation you need
2. Unigrams, bigrams and trigrams: construct them correctly
3. Collocations and the purpose of a language model
4. Probability chain rule and the Markov assumption