Groups
Category
VC dimension measures how many distinct labelings a hypothesis class can realize on any set of points of a given size.
The biasโvariance tradeoff explains how prediction error splits into bias squared, variance, and irreducible noise for squared loss.