Overfitting

Memorising versus learning

An interactive lesson with 6 steps, about 5 minutes: you predict, drag and solve, with instant feedback on every answer. It runs in your browser and needs JavaScript.

What you’ll learn

• Overfitting: fitting the noise in training data rather than the underlying pattern.
• Training error alone is misleading; judge models on held-out test data.
• More data, simpler models and regularisation reduce overfitting.

Steps

1. Too flexible
2. Spotting it
3. Why degree 7?
4. Regularisation
5. The test set
6. Early stopping