How does SVM relate to risk minimization in statistical learning theory?

SVM minimizes structural risk by balancing training error and model complexity.

SVM is grounded in Structural Risk Minimization (SRM), which extends Empirical Risk Minimization. Instead of only minimizing training error, SVM also controls model complexity via margin maximization. This reduces overfitting and improves generalization bounds.