Divyesh Narayanan

@sylvchev when you get the time, if you can check out the recent issues I raised that came from running the benchmark pipeline and possibly the Colab notebook itself which...

 I think the evaluation schemes that allow doing these as explained in the picture above would be a good addition for MOABB

@sylvchev I think the combination of FB and SinglePass results is also done [Line](https://github.com/NeuroTechX/moabb/pull/264/files#diff-5e3ef2e09eb95efd111d15f0abc9e76700075796d7deb46bdf883c19dda0d646R141) If I remember correctly this shouldn't be done in case of MI.

Hey @GillesVandewiele @rtavenar I was trying out using LearningShapelets on a variable length time series data and ran into this error too.. I got nans right from the first epoch....

> henrydeclety, I think the length is not a bug but expected behaviour. > > If you read the docstring on how the algorithm works, this is the same as...

Quick question @fkiraly . What is your general approach while doing these sorts of tracebacks? I generally use PyCharm and I'm trying to see if I can set a breakpoint...

Never mind that question. I managed to trace the feature_importances_ being called. If its a ComposableTimeSeriesForestClassifier that's called, then this version of feature_importances_ is called https://github.com/sktime/sktime/blob/c886759acbd79b008165604cd53920ca608c0683/sktime/series_as_features/base/estimators/_ensemble.py#L289 Otherwise, if the TimesSeriesForestClassifier...

I've added an example comparing spectral_connectivity_epochs and time using simulated data and coherence here. Haven't yet cleared out some of the testing code. - Could have a short part showing...

Hey @adam2392 , no issues! I can check once again.. I have been having some doubts about the indices parameter as well. I tried using specified channel indices to calculate...

@adam2392 While I didn't look specifically into the exact working, I did further testing with the methods. These are my testings for the single epoch over time methods - Coherence...