This past post gives a technical overview of how to adjust your probabilities in the ML context when under-sampling. Recently gave a talk at the Ai4 Finance conference, where I tried to give more context around the business use case. Here’s the updated notebook.

Presentation PDF