Python bindings to the FEAST Feature Selection Toolbox..

Downlaod Version 1.1

PyFeast is a interface for the FEAST feature selection toolbox, which was originally written in C with a interface to Matlab.

Because Python is also commonly used in computational science, writing bindings to enable researchers to utilize these feature selection algorithms in Python was only natural.

At Drexel University's EESI Lab, we are using PyFeast to create a feature selection tool for the Department of Energy's upcoming KBase platform. We are also integrating a tool that utilizes PyFeast as a script for Qiime users: Qiime Fizzy Branch

In order to use the feast module, you will need the following dependencies

• Python 2.7
• Numpy
• Linux or OS X
• MIToolbox
• FEAST v1.1.1 or higher

python ./setup.py build
sudo python ./setup.py install

See test/test.py for an example with uniform data and an image data set. The image data set was collected from the digits example in the Scikits-Learn toolbox. Make sure that if you are loading the data from a file and converting the data to a numpy array that you set order="F". This is very important.

We have documentation for each of the functions available here

• FEAST - The Feature Selection Toolbox
• Fizzy - A KBase Service for Feature Selection
• Conditional Likelihood Maximisation: A Unifying Framework for Information Theoretic Feature Selection