pyriemann.utils.mean.mean_covariance

pyriemann.utils.mean.mean_covariance(covmats, metric='riemann', sample_weight=None, *args)[source]

Return the mean covariance matrix according to the metric

Parameters:
  • covmats – Covariance matrices set, Ntrials X Nchannels X Nchannels
  • metric – the metric (Default value ‘riemann’), can be : ‘riemann’ , ‘logeuclid’ , ‘euclid’ , ‘logdet’, ‘indentity’, ‘wasserstein’, ‘ale’,

‘harmonic’, ‘kullback_sym’ :param sample_weight: the weight of each sample :param args: the argument passed to the sub function :returns: the mean covariance matrix