Imbalanced class handler
Created by: aPovidlo
Adding class Imbalanced handler to FEDOT. Was used simple implementation from sklearn.utils to classification with two classes. For example, target can be a numer from {0, 1}. Elements in class 0 more than class 1. To balance them, one of two strategies is chosen: expand_minority or reduce_majority. For expand_minority elements of class 1 are expanding to n_samples or to the shape of class 0. In case of reduce_majority elements of class 0 are reduce to n_samples or to the shape of class 1. Also it is possible to specify boolean param replace which means that implements resampling with replacement. If False, this will implement (sliced) random permutations.