[WIP] Windowed transformation and Bearing Fault example

[lucianolorenti]

This PR includes some additional transformer, in particular, the class SplitIntoWindows. This class splits the window into multiple windows. Unlike using sliding windows, this allows you to transform each window independently.

An example of using this class is provided. The example tries to be similar to the one presented here [1]. The MFPT dataset is also included in the library.
The example basically:

• splits the signals into windows
• performs a band-pass filter whose frequency bands are obtained through a kurtogram (Code for doing the kurtogram is also added)
• Compute the envelope of the filtered signal
• Extract features from the envelope
• Classify the features extracted in each window, intro three classes: NORMAL, OUTER_RACE_FAULT, INNER_RACE_FAULT

[1] https://it.mathworks.com/help/predmaint/ug/Rolling-Element-Bearing-Fault-Diagnosis.html