Synthetic Minority Oversampling Technique (SMOTE) is an approach used for dealing with imbalanced datasets before running them through machine learning models.