Unbalanced Datasets

Multiple methods to deal with unbalanced datasets can be used to do oversampling, under-sampling or a mixture of both (Random, SMOTE, SMOTEENN, SMOTETomek and ClusterCentroids).

from deepmol.imbalanced_learn.imbalanced_learn import SMOTEENN

train_dataset = SMOTEENN().sample(train_dataset)