This dataset came from kaggle. The dataset has two classes which were both quite unbalanced, this .py file does some data pre-processing, smoothing to remove signal noise, and then uses a combination of undersampling, oversampling, and data synthesization to balance the classes to train a model.