Explanation:
Both ridge regression and LASSO regression decrease the variance in exchange for a small increase in bias. Ridge regression is preferred when the response is a function of many predictors (i.e., shrinks some coefficients closer to zero). LASSO regression is preferred when the response can be adequately explained by only a subset of predictors (i.e., sets some coefficients to exactly zero).
(Book 2, Module 26.1, LO 26.d)
Ultimate access to all questions.
Question 44
When comparing ridge regression with LASSO regression, which of the following statements is true?
A
Ridge regression results in reduced variance but increased bias.
B
LASSO regression results in both reduced variance and reduced bias.
C
LASSO regression is preferred if the response can be explained by many predictors.
D
Ridge regression is preferred if the response can be explained by a subset of predictors.
No comments yet.