Definition #
- Overfitting: Model memorizes training data but fails on new data.
- Underfitting: Model is too simple to capture patterns.
Key Characteristics #
- High variance (overfitting)
- High bias (underfitting)
Why It Matters #
Proper balancing improves real-world accuracy by 25-40%.
Solutions #
- Regularization (for overfitting)
- Feature engineering (for underfitting)