pyriemann is a python package for covariance matrices manipulation and classification through riemannian geometry.
The primary target is classification of multivariate biosignals, like EEG, MEG or EMG.
This is work in progress ... stay tuned.
This code is BSD-licenced (3 clause).
The documentation is available on http://pyriemann.readthedocs.io/en/latest/
pip install pyriemann
or using pip+git for the latest version of the code :
pip install git+https://github.com/alexandrebarachant/pyRiemann
Anaconda is not currently supported, if you want to use anaconda, you need to create a virtual environment in anaconda, activate it and use the above command to install it.
For the latest version, you can install the package from the sources using the setup.py script
python setup.py install
or in developer mode to be able to modify the sources.
python setup.py develop
How to use it
Most of the functions mimic the scikit-learn API, and therefore can be directly used with sklearn. For example, for cross-validation classification of EEG signal using the MDM algorithm described in  , it is easy as :
import pyriemann from sklearn.model_selection import cross_val_score # load your data X = ... # your EEG data, in format Ntrials x Nchannels X Nsamples y = ... # the labels # estimate covariances matrices cov = pyriemann.estimation.Covariances().fit_transform(X) # cross validation mdm = pyriemann.classification.MDM() accuracy = cross_val_score(mdm, cov, y) print(accuracy.mean())
You can also pipeline methods using sklearn Pipeline framework. For example, to classify EEG signal using a SVM classifier in the tangent space, described in  :
from pyriemann.estimation import Covariances from pyriemann.tangentspace import TangentSpace from sklearn.pipeline import make_pipeline from sklearn.svm import SVC from sklearn.model_selection import cross_val_score # load your data X = ... # your EEG data, in format Ntrials x Nchannels X Nsamples y = ... # the labels # build your pipeline covest = Covariances() ts = TangentSpace() svc = SVC(kernel='linear') clf = make_pipeline(covest,ts,svc) # cross validation accuracy = cross_val_score(clf, X, y) print(accuracy.mean())
Check out the example folder for more examples !
If you make a modification, run the test suite before submitting a pull request
 A. Barachant, M. Congedo ,"A Plug&Play P300 BCI Using Information Geometry", arXiv:1409.0107. link
 M. Congedo, A. Barachant, A. Andreev ,"A New generation of Brain-Computer Interface Based on Riemannian Geometry", arXiv: 1310.8115. link
 A. Barachant and S. Bonnet, "Channel selection procedure using riemannian distance for BCI applications," in 2011 5th International IEEE/EMBS Conference on Neural Engineering (NER), 2011, 348-351. pdf
 A. Barachant, S. Bonnet, M. Congedo and C. Jutten, “Multiclass Brain-Computer Interface Classification by Riemannian Geometry,” in IEEE Transactions on Biomedical Engineering, vol. 59, no. 4, p. 920-928, 2012. pdf
 A. Barachant, S. Bonnet, M. Congedo and C. Jutten, “Classification of covariance matrices using a Riemannian-based kernel for BCI applications“, in NeuroComputing, vol. 112, p. 172-178, 2013. pdf
- Updated for scikit-learn v0.22
- Remove support for python 2.7
- Update the examples to take into account last version of MNE
- Fix cospectrum and coherency estimation.
- Added a permutation test for generic scikit-learn estimator
- Stats module refactoring, with distance based t-test and f-test
- Removed two way permutation test
- Support for python 3.5 and 3.6
- Improved documentation
- Added TSclassifier for out-of the box tangent space classification.
- Added Wasserstein distance and mean.
- Added NearestNeighbor classifier.
- Added Softmax probabilities for MDM.
- Added CSP for covariance matrices.
- Added Approximate Joint diagonalization algorithms (JADE, PHAM, UWEDGE).
- Added ALE mean.
- Added Multiclass CSP.
- API: param name changes in
CospCovariancesto comply to Scikit-Learn.
- API: attributes name changes in most modules to comply to the Scikit-Learn naming convention.
- Added Harmonic mean
- Added Kullback leibler mean
- Added multiprocessing for MDM with joblib
- Added kullback-leibler divergence
- Added Riemannian Potato
- Added sample_weight for mean estimation and MDM