Techniques to address imbalanced data in the context of classification, while keeping the data distribution in mind.