Given a sentence, the tagger assigns each word a part-of-speech tag (noun, verb, adjective, ...) using a Hidden Markov Model: transition probabilities model how tags follow each other, emission ...