Why SMOTE is not necessarily the answer to your imbalanced dataset

Everybody loves SMOTE, but is it really a silver bullet?

When the topic of imbalanced classification comes up, many people praise SMOTE as the go-to approach. And indeed, popular algorithms have survived some evolutionary selection process by the community. However, we should still not just blindly throw canned solutions at our problems. To prove this point, we can construct actual toy examples that make SMOTE fail miserably. Before we do that, let us …

dataset editors pick smote