pyriemann.utils.mean.mean_ale

pyriemann.utils.mean.mean_ale(covmats, tol=1e-06, maxiter=50, sample_weight=None)[source]

Return the mean covariance matrix according using the AJD-based log-Euclidean Mean (ALE). See [1].

Parameters:
  • covmats – Covariance matrices set, Ntrials X Nchannels X Nchannels
  • tol – the tolerance to stop the gradient descent
  • maxiter – The maximum number of iteration, default 50
  • sample_weight – the weight of each sample
Returns:

the mean covariance matrix

Notes

New in version 0.2.4.

References

[1] M. Congedo, B. Afsari, A. Barachant, M. Moakher, ‘Approximate Joint Diagonalization and Geometric Mean of Symmetric Positive Definite Matrices’, PLoS ONE, 2015