moabb.pipelines.utils.FilterBank#

moabb.pipelines.utils.FilterBank(estimator, flatten=True)[source]#

Apply a given identical pipeline over a bank of filter.

The pipeline provided with the constrictor will be appield on the 4th axis of the input data. This pipeline should be used with a FilterBank paradigm.

This can be used to build a filterbank CSP, for example:

pipeline = make_pipeline(FilterBank(estimator=CSP()), LDA())
Parameters:

  • estimator (sklean Estimator) – the sklearn pipeline to apply on each band of the filter bank.

  • flatten (bool (True)) – If True, output of each band are concatenated together on the feature axis. if False, output are stacked.