Chestnut Machine Learning Suite

Version 0.1.1

by Damian Eads¹, Karen Glocer¹, Andy Fraser², and Edward Rosten³

University of California, Santa Cruz; Santa Cruz, CA USA
Los Alamos National Laboratory; Los Alamos, NM USA
University of Cambridge; Cambridge, UK

About

The Chestnut Machine Learning Suite is a collection of machine learning algorithms written in Python with some code written in C for efficiency. Most algorithms are called with a simple, functional API with input data encoded as arrays. Some packages have a MATLAB-like API to enable migration from a MATLAB environment. The class hierarchy is minimal to enable new users to quickly learn and call Chestnut code from their own codes. The first alpha release includes the following packages:

boosting: AdaBoost, LPBoost, BrownBoost, MADABoost, SoftBoost, TotalBoost (some boosters require CVXOPT)
hmm: hidden markov models with support for discrete and continuous emission distributions and multiple training sequences.
cluster: hierarchical, k-means, QT, and shifting means clustering algorithms
knn: Voronoi tesselations, kd-trees, k-nearest neighbor classifiers
linear: basic Fisher's linear discriminant analysis

Documentation

We are in the process of reformatting the documentation for Sphinx. In the mean time, we recommend using the help(name) with IPython to get detailed help with Chestnut packages, modules, and functions.

Download

You can download a very preliminary alpha release here.

License

GNU General Public License (GPL) Version 3.0