FsHiddenMarkov

Hidden Markov Model for F#. The implementation is based on a paper by Mark Stamp.

A model is defined as $ \lambda = (A, B, \pi) $ with:

• $ A = N x N $ state transition matrix, $ a_{ij} $ is the probability of transitioning from state $ i $ to state $ j $.
• $ B = N x M $ observation probability matrix, $ b_{j}(k) $ being the probability of observing $ k $ in state $ j $.
• $ \pi = N x 1 $ is the initial state distribution vector.

States and observations are simply integers. The mapping to domain types is kept outside of this project.

Features

1. Compute the probability of an observation sequence, given a model.
2. Predicts the most likely hidden state sequence, given observations.
3. Estimates a Hidden Markov Model given an initialization and observational data.

References

• Stamp, Mark - A Revealing Introduction to Hidden Markov Models (2021)