Continuous-time Hidden Markov Model
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
docs
hmms
test
.gitignore
LICENSE
MANIFEST.in
README
README.md
README.rst
dev
hmms.ipynb
requirements.txt
script.sh
setup.cfg
setup.py

README.md

Intro

HMMs is the Hidden Markov Models library for Python. It is easy to use, general purpose library, implementing all the important submethods, needed for the training, examining and experimenting with the data models.

The effectivness of the computationally expensive parts is powered by Cython.

You can build two models:

Discrete-time Hidden Markov Model
Usually just reffered as the Hidden Markov Model.
Continuous-time Hidden Markov Model
The variant of the Hidden Markov Model, where the state transition can occure in the continuous time, and that allows random distribution of the observation times.

Before starting to work, it is recommended to go trough tutorial with examples, the ipython notebook, covering most of the main usecases.

For deeper understanding of the topic you can see the corresponding diploma thesis. Or read the main referenced articles: Dt-HMM, Ct-HMM .

Requirements

  • python 3.5
  • libraries: Cython, ipython, matplotlib, notebook, numpy, pandas, scipy,
  • libraries for testing environment: pytest

Download & Install

After installing Numpy and Cython, you can install the package directly from test pypi.

(env)$ python -m pip install numpy cython
(env)$ python -m pip install hmms