Overfitting — Glossary

Overfitting

Concept

overfitting

When a model memorises training noise and fails on new data. Held-out test sets, cross-validation, and tuned hyperparameters guard against it.

Train R² 0.99 but test R² 0.3 → overfit.