Hidden Markov Models are powerful statistical models used for sequential data analysis. A multivariate hidden Markov model extends the traditional HMM to handle multiple observations at each state. In this repo, I have implemented several functions and algorithms in HMM for multivariate cases such as alpha recursion, beta recursion, pair marginals, expectation maximization, and Viterbi.
more details can be found in the notebook, I recommend opening the notebook in Google Colab.
the report is in Persian.