How do precision and recall apply to a fraud detection model, and which one would you prioritize for high-value transactions?
