Structural Risk Minimization

A classical statistical learning framework introduced by Vapnik that trades off empirical risk against model complexity. It selects the model with the best performance at a given complexity level, providing theoretical control over overfitting through capacity-based generalization bounds.