Real-world uses

Created for this library

1. 1.

   A fraud team handles an imbalanced dataset with class weights so the rare positive class drives more of the loss during training.

2. 2.

   A medical screening team handles an imbalanced dataset for rare diseases by reporting precision-recall curves rather than ROC.

3. 3.

   A churn team handles an imbalanced dataset by undersampling retained customers during training to balance the gradient signal.