=============================== mca

.. image:: https://badge.fury.io/py/mca.png :target: http://badge.fury.io/py/mca

.. image:: https://travis-ci.org/esafak/mca.png?branch=master :target: https://travis-ci.org/esafak/mca

mca is a Multiple Correspondence Analysis <http://en.wikipedia.org/wiki/Multiple_correspondence_analysis>_ (MCA) package for python, intended to be used with pandas <http://pandas.pydata.org/>. MCA is a feature extraction <http://en.wikipedia.org/wiki/Feature_extraction> method; essentially PCA <http://en.wikipedia.org/wiki/Principal_component_analysis>_ for categorical variables <http://en.wikipedia.org/wiki/Categorical_variable>. You can use it, for example, to address multicollinearity <http://en.wikipedia.org/wiki/Multicollinearity> or the curse of dimensionality <http://en.wikipedia.org/wiki/Curse_of_dimensionality>_ with big categorical variables.

Installation

.. code :: bash

pip install --user mca

Usage

Please refer to the usage notes <https://github.com/esafak/mca/blob/master/docs/usage.rst>_ and this illustrated ipython notebook <http://nbviewer.ipython.org/github/esafak/mca/blob/master/docs/mca-BurgundiesExample.ipynb>_.

Reference

Michael Greenacre, Jörg Blasius (2006). Multiple Correspondence Analysis and Related Methods <http://www.crcpress.com/product/isbn/9781584886280>_, CRC Press. ISBN 1584886285.