Bayesian_HMM

This project is the experimental part of my undergraduate thesis.

The project includes the things below:

• The main part is in the main.py which can set the hyper-parameters.
• The basic_algorithm.py includes NIX posterior estimation (VB algorithm is a single function), forward-backward algorithm and Viterbi algorithm. The skills of calculation refers part of hmmlearn (https://github.com/hmmlearn/hmmlearn).
• The core_model.py includes the four deterministic algorithms introduced by the paper "Estimation of Viterbi path in Bayesian hidden Markov models".
• The data_generate.py includes the generation of initial sequences and TMC model data introduced by the paper "Triplet Markov Chains in hidden signal restoration".