Hidden Markov Model (HMM)

A model when the actual states of a Markov chain is hidden or unobservable.

Is a double stochastic process.

• The hidden (actual) states of the Markov process
• These hidden states produce observations based on a probability distribution $P(O_t|S_t)$

Note the observations only depend on the current state (in the basic model)

---

References

@sucar2020 - Chapter 5 Probabilistic Graphical Model (PGM)