pyGPs is a library containing an object-oriented python implementation for Gaussian Process (GP) regression and classification.
Clone or download
Pull request Compare This branch is 62 commits behind marionmari:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
doc
pyGPs
CHANGELOG.txt
COPYRIGHT.txt
FUNCTIONALITY.txt
MANIFEST.in
README.md
pyGPs_API.pdf
setup.py

README.md

================================================================================ Marion Neumann [marion dot neumann at uni-bonn dot de] Daniel Marthaler [dan dot marthaler at gmail dot com] Shan Huang [schan dot huang at gmail dot com] Kristian Kersting [kristian dot kersting at cs dot tu-dortmund dot de]

This file is part of pyGPs.
The software package is released under the BSD 2-Clause (FreeBSD) License.

Copyright (c) by
Marion Neumann, Daniel Marthaler, Shan Huang & Kristian Kersting, 18/02/2014

================================================================================

pyGPs is a library containing code for Gaussian Process (GP) Regression and Classification.

Here is the online documentation: ONLINE documentation.

pyGPs is an object-oriented implementation of GPs. Its functionality follows roughly the gpml matlab implementation by Carl Edward Rasmussen and Hannes Nickisch (Copyright (c) by Carl Edward Rasmussen and Hannes Nickisch, 2013-01-21).

Standard GP regression and (binary) classification as well as FITC (sparse GPs) inference is implemented. For a list of implemented covariance, mean, likelihood, and inference functions see list_of_functions.txt. The current implementation is optimized and tested, however, the work on this library is still in progress. We appreciate any feedback.

A comprehensive introduction to functionalities and demonstrations can be found in the doc folder; just open /doc/build/html/index.html in your browser to get to the html documentation of the whole package.

Further, pyGPs includes implementations of

  • minimize.py implemented in python by Roland Memisevic 2008, following minimize.m which is copyright (C) 1999 - 2006, Carl Edward Rasmussen
  • scg.py (Copyright (c) Ian T Nabney (1996-2001))
  • brentmin.py (Copyright (c) by Hannes Nickisch 2010-01-10.)

Installing pyGPs

Download the archive and extract it to any local directory.

You can either add the local directory to your PYTHONPATH:

export PYTHONPATH=$PYTHONPATH:/path/to/local/directory/../parent_folder_of_pyGPs

or install the package using setup.py:

python setup.py install

or install via pip::

pip install pyGPs 

Requirements

  • python 2.6 or 2.7
  • scipy (v0.13.0 or later), numpy, and matplotlib: open-source packages for scientific computing using the Python programming language.

Acknowledgements

The following persons helped to improve this software: Roman Garnett, Maciej Kurek, Hannes Nickisch, Zhao Xu, and Alejandro Molina.

This work is partly supported by the Fraunhofer ATTRACT fellowship STREAM.