Definition #
Assessing how well a trained model generalizes to unseen data.
Key Characteristics #
- Train/test splits
- Cross-validation
- Confusion matrices
Why It Matters #
Prevents deployment of flawed models.
Examples #
- Scikit-learn’s
train_test_split - K-fold validation