Can online learning or ReservoirPy in general be used for time series classification?

[jagandecapri]

Hi,

Thank you for this library.

Currently, I am trying to use online learning as demonstrated in https://github.com/reservoirpy/reservoirpy/blob/master/tutorials/Online learning/online_learning_example_MackeyGlass.py for a classification task. The input dataset is of the shape (batch, time_sequence, features) and for each batch there is one prediction (0 or 1) which yields the shape (batch, prediction).

As I was playing around the online learning example (I could be wrong), it seems to me that online learning in ReservoirPy is built for time series forecasting where the predictions (teachers for training) should be (batch, time_sequence, prediction) or in other words, each time_sequence in each batch has its prediction.

Hence, my questions are;

1. Can online learning be used for time series classification as I outlined above?
2. Generally, can ReservoirPy be used for time series classification as I outlined above?