Neurobot » Multivariate Pattern Analysis in Python

26
Feb
Multivariate Pattern Analysis in Python

PyMVPA is a Python module intended to ease multivariate pattern classification analyses of large datasets.

In the neuroimaging contexts such analysis techniques are also known as decoding or MVPA analysis.
http://www.pymvpa.org
PyMVPA:
* provides high-level abstraction of typical processing steps and a
number of implementations of some popular statistical learning
algorithms, such as
- classifiers: SVM, kNN, SMLR, etc
- feature selection methods: RFE, IFS, etc
* is not limited to the neuroimaging domain, but is eminently suited
for such datasets (e.g. transparent I/O for Nifti/Analyze data
formats)
* allows researchers to compress complex analyses into a small amount
of code. This makes it possible to complement publications with the
source code, which leads to an increase in scientific progress due
to the superior accessibility of information and reproducibility of
scientific results
* is truly a free software (MIT license) and additionally
requires nothing but free-software to run
* is fully or partially supported on any platform supported by Python
(depending on the availability of optional external dependencies)
* provides high-level “house-keeping” functionality done by the base
classes, reducing the necessary amount of code needed to contribute
a new fully-functional algorithm
* contains extensive user-manual with concrete and tested examples of the analysis pipelines
PyMVPA is a collaborative project, which was initiated by Michael Hanke and Yaroslav O. Halchenko. It welcomes new contributors and users. All the source materials (code, manuals, website) are available for the collaborative development from a distributed version control system (git).
The PyMVPA developers team currently consists of:
* Michael Hanke, University of Magdeburg, Germany
* Yaroslav O. Halchenko, Rutgers University Newark, USA
* Per B. Sederberg, Princeton University, USA
* Emanuele Olivetti, University of Trento, Italy