3) Explain how k-fold cross-validation is implemented.

4) What are the advantages and disadvantages of k-fold cross- validation relative to: The validation set approach?

5) What are principal components? Is this a supervised or unsupervised learning method?

6) What is the main difference between ridge and lasso regression?

7) What happens to R-squared as you continue to add features to your model?

8) How does one select the correct Tuning Parameter ??

9) Under what circumstances is lasso more appropriate than ridge regression?

10) We perform best subset, forward stepwise, and backward stepwise selection on a single data set. For each approach, we obtain p + 1 models, containing 0, 1, 2, . . . , p predictors. Which of the three models with k predictors has the smallest test RSS?

31) Which fit is better, and why?

> #Validation set - Linear > Im. fit1 mean((logtdc-predict(Im. fit1, ceo_rproject)) [-train]^2) [1] 1. 024548 > #Validation set - Quadratic fit > Im. fit2 mean((logtdc-predict(lm. fit2, ceo_rproject))[-train]^2) [1] 0. 9988668\f