ML Knowledge

What is the difference between Ridge and Lasso?

Data Scientist

Chime

Sprinklr

Benchling

ServiceNow

SurveyMonkey

Segment

Did you come across this question in an interview?

Answers

Anonymous

4Strong
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.
  • What is the difference between Ridge and Lasso?
  • In what ways do Ridge and Lasso diverge from each other?
  • How different are Ridge and Lasso, and in what regards?
  • What are the distinctions between Ridge and Lasso?
  • Do you consider Ridge and Lasso to be dissimilar, and if so, in what respect?
  • What sets Ridge and Lasso apart from each other?
  • Can you elucidate how Ridge and Lasso contrast?
  • What are the differences between Ridge and Lasso, and how do they manifest?
  • Could you discuss the dissimilarities between Ridge and Lasso?
  • Could you elaborate on the contrast between Ridge and Lasso?
Try Our AI Interviewer

Prepare for success with realistic, role-specific interview simulations.

Try AI Interview Now

Interview question asked to Data Scientists interviewing at Peloton, Adyen, Careem and others: What is the difference between Ridge and Lasso?.