# Baum-Welch Algorithm

A Baum-Welch Algorithm is an Expectation-Maximization (EM) algorithm that can fit a Hidden Markov Model.

## References

### 2015b

1. Compute $Q(\theta, \theta_s ) = \sum_{z\in Z} \log [P(X , z; \theta)] P(z|X ; \theta_s )$.
2. Set $\theta^{s+1} = \underset{\theta}{\operatorname{arg\,max}} \,Q(\theta, \theta_s )$