Can you elucidate how Ridge and Lasso contrast?

Question

Anonymous · Accepted Answer

Lasso and Ridge are both regularization methods which adds penalties to the loss
function in linear regression. Ridge uses L2 norm which is the sum of the
squared of the coef to penalize the coefficient while Lasso uses L1 which is the
sum of the absoulte values of coefficients to penalize.  Lasso can shrink
coefficients towards 0 to perform feature selection while Ridge cannot. Ridge is
best to use when your predictors has some collinearity, though some weak
predictors but you still want to keep them all. Lasso is mainly used for feature
selection, it's best to use when you have a lot of predictiors and you suspect
that some of them are not relevant.

Can you elucidate how Ridge and Lasso contrast?

Practice this question with AI

Go Premium

Practice More Questions

Community Answers

Unlock Community Insights

Turn Adequate Answers Into Great Ones