gnumpi / t-student-mixture-models

Contribution: addresses numerical stability issues of original implementation. Moves calculations to log-space. Branch: log_space

This fork addresses numerical stability issues of the original implementation. Calculation is moved to log-space.

t-Student-Mixture-Models

Implementation of the paper: 'Robust mixture modelling using the t distribution', D. Peel and G. J. McLachlan.

• Only Python >= 3.7 supported.
• Code coverage: 76%.

Install with pip

$ python3 -m pip install smm --user

Install from source

$ git clone https://github.com/luiscarlosgph/t-Student-Mixture-Models.git
$ cd t-Student-Mixture-Models
$ python3 setup.py install --user

Usage

See example in src/example.py.

$ python3 src/example.py

Unit tests

To run the tests execute:

$ python3 setup.py test

Coverage

To run the coverage test:

$ python3 -m pip install coverage
$ python3 -m coverage run setup.py test
$ python3 -m coverage html

Then open 'htmlcov/index.html' and search for the line containing 'smm/smm.py'.

Documentation

See t-Student-Mixture-Models documentation.

Author

Luis Carlos Garcia-Peraza Herrera ([email protected]).

License

BSD 3-Clause License, see LICENSE file for more information.