Lecture
This lecture delves into the concept of convexity in the ROC plane, explaining how randomized classifiers achieve convex combinations. It provides a detailed proof that the FPR of a randomized classifier equals a convex combination of its base classifiers' FPRs.