PyFeast
Python bindings to the FEAST Feature Selection Toolbox.
About PyFeast
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
Requirements
In order to use the feast module, you will need the following dependencies
- Python 2.7
- Numpy
- Linux or OS X
Installation
To install the FEAST interface, you'll need to build and install the FEAST libraries first, and then install python.
Make MIToolbox and install it:
cd FEAST/MIToolbox
make
sudo make install
Make FSToolbox and install it:
cd FEAST/FSToolbox
make
sudo make install
Run ldconfig to update your library cache:
sudo ldconfig
Install our PyFeast module:
python ./setup.py build
sudo python ./setup.py install
Demonstration
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.
Documentation
We have documentation for each of the functions available here