Marion Neumann [marion dot neumann at uni-bonn dot de] Daniel Marthaler [marthaler at ge dot com] Shan Huang [shan dot huang at iais dot fraunhofer dot de] 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.
pyGPs is an object-oriented implementation of GPs. Its functionalities follow roughly the gpml matlab implementaion 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 (spares 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.
For 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.)
Download the archive and extract it to any local directory. Add the local directory to your PYTHONPATH: export PYTHONPATH=$PYTHONPATH:/path/to/local/directory/ (parent directory of pyGPs)
- python 2.6 or 2.7
- scipy, numpy, and matplotlib: open-source packages for scientific computing using the Python programming language.
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.