README

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 <https://github.com/lopatovsky/CT-HMM/blob/master/hmms.ipynb>`__,
covering most of the main usecases.

For **deeper understanding** of the topic you can see the corresponding
`diploma thesis <https://github.com/lopatovsky/DP>`__. Or read the
main referenced articles:
`Dt-HMM <http://www.ece.ucsb.edu/Faculty/Rabiner/ece259/Reprints/tutorial%20on%20hmm%20and%20applications.pdf%3E>`__,
`Ct-HMM <https://web.engr.oregonstate.edu/~lif/nips2015_CTHMM_learning_camera_ready.pdf>`__
.

-  Sources of the project:
   `Pypi <https://pypi.python.org/pypi/hmms>`__,
   `Github <https://github.com/lopatovsky/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 pypi.

::

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