You have a classification model where one class is much rarer than the other, and plain accuracy is not telling you much. The team wants to know whether the model is actually useful for the rare class.
How would you evaluate a classification model with imbalanced classes?
