Overfitting vs. Underfitting

Overfitting occurs when a model learns training data too well but fails on new data; underfitting when it fails to capture patterns.