← Quiz: Regularisation (Ridge and Lasso): all 10 questions
Quiz: Regularisation (Ridge and Lasso) · Question 10 of 10
The degree-9 model with λ = 100 had high training AND validation error. Why?
The degree-9 model with λ = 100 had high training AND validation error. Why?
- A.It was still overfitting
- B.The data had no noise
- C.The penalty was so strong that the weights were squashed near zero and the model underfit
- D.Ridge cannot be used with polynomial features
Answer & explanation
Correct answer: The penalty was so strong that the weights were squashed near zero and the model underfit
Too much regularisation brings back high bias.
Can you get the rest right without peeking?
Take the timed Quiz: Regularisation (Ridge and Lasso) quiz free. Pass it to earn XP and count towards your certificate.
Start the quiz →