scikit-cmeans is a (currently rather small) package designed to facilitate exploration of fuzzy clustering algorithms in a way that is both readily usable and easily extended. The API is loosely based on that of scikit-learn.
Many packages already distribute versions of the C-means algorithm, but fuzzy clustering is a rich field. It is often necessary or desirable to explore different parameters or methods rapidly, or develop small tweaks to algorithms without having to rewrite the entire codebase. This project aims to make that as easy as possible, by providing a simple API promoting small, decoupled, connectable tools.
To install the latest version of scikit-cmeans, you can use pip
to install
from the main GitHub repository.
You must ensure that you have pip
installed. See
here for more information.
pip install git+git://github.com/bm424/scikit-cmeans.git@master
Warning
Installation has not yet thoroughly tested.
Out-of-the-box, scikit-cmeans provides algorithms for hard clustering, probabilistic and possibilistic c-means clustering, and a plugin for the Gustafson-Kessel variant for ellipsoidal clusters. Any of the basic algorithms can be used with any distance metric available from scipy, or use a custom distance function. Data of any dimensionality is supported