ISLP Ch 06 Quiz

Suppose we have 5 predictors in a model. According to best subset selection method how many models we have to consider?

• 64
• 32
• 16
• None of the above

Suppose we have 5 predictors in a model. According to forward stepwise selection method how many models we have to consider?

• 64
• 32
• 16
• None of the above

Suppose we have 5 predictors in the model and let's assume M2 is the best model when k=2 in best subset selection method. From how many models M2 was selected as the best?

• 10
• 20
• 30
• None of the above

Suppose we have 5 predictors in the model. Let's assume M2 is the best model among all the models with two predictors i.e. when k=2 in best subset selection method. M2 was selected as the best model based on which measure?

• Smallest RSS
• R^2
• Adjusted R^2
• AIC
• BIC

Suppose we have 5 predictors in the model. Let's assume M3 is the best model among all the models with all predictors in best subset selection method. Which of the following measure(s) was NOT used to select M3 as the best model

• Smallest RSS
• R^2
• Adjusted R^2
• AIC
• BIC

Suppose we have 4 predictors (x1,x2,x3 and x4) in a model and we are considering forward stepwise selection method. When k=2, we found that the model with x2 and x3 performs the best. From how many models this best model was chosen?

• 3
• 4
• 10
• 20
• None of the above

Suppose we have 4 predictors (x1,x2,x3 and x4) in a model and we are considering backward stepwise selection method. When k=3, we found that the model with x1 , x3 and x4 performs the best. From how many models this best model was chosen?

• 3
• 4
• 10
• 20
• None of the above

Which of the following is NOT true about forward and backward stepwise selection?

• Like forward stepwise selection, the backward selection approach searches through only 1+p(p+ 1)/2 models
• Backward selection requires that n should be higher than p
• Forward selection requires that n should be higher than p
• Forward selection does not require that n should be higher than p
• None of the above

Which of the following is true about shrinkage methods?

• In this method all the predictors are used
• It constraints or regularizes the coefficient estimates
• It shrinks the coefficient estimates to zero
• All of the above
• None of the above

In ridge regression, the following mathematical expression is minimized to find the coefficient estimates. Which of the following is true about this expression?

• When 𝝀=0, all the coefficient estimates are also zero.
• When λ → ∞, the shrinkage penalty reduces to zero.
• 𝝀 controls the relative importance of the two terms in above expression
• All of the above
• None of the above