Natural Language Processing · TEE · formulas-chapter

HMM & decoding

Formulas explained through full worked examples

1. Markov chains: read a transition table and multiply a path
2. Estimate transition and emission probabilities from tagged data
3. HMM path probabilities: three complete numerical examples
4. POS HMM worked corpus: Mary, Jane, can, will and spot
5. Why exhaustive search grows exponentially
6. Viterbi decoding: compare incoming paths, keep the best, trace back
7. A path score is not the probability of the observed sentence