Given a scenario where the feature correlation is very high, how would you choose between Ridge, Lasso, and ElasticNet?

  • It doesn't matter
  • Prefer ElasticNet
  • Prefer Lasso
  • Prefer Ridge
ElasticNet is preferred when there's multicollinearity, as it combines L1 and L2 penalties, balancing the properties of Ridge and Lasso.
Add your answer
Loading...

Leave a comment

Your email address will not be published. Required fields are marked *