Add feature names in DataSet

[LouisCarpentier42]

The DataSet stores the time series data as a set of numpy arrays (separated in train/test data + labels of each set). In several cases, this approach loses information, as you do not know from which real-world process the data originates.

This can be fixed by adding a list of strings to the DataSet, which give the name of the feature that is present in the data.

[LouisCarpentier42]

In addition to the feature names, it would also be useful to store the times of each observation.