You built a model using Lasso regularization but some important features were wrongly set to zero. How would you modify your approach to keep these features?

  • Combine with ElasticNet
  • Decrease L1 penalty
  • Increase L1 penalty
  • Switch to Ridge
Combining with ElasticNet allows for balancing between L1 and L2 penalties, thus avoiding complete elimination of important features by the L1 penalty.
Add your answer
Loading...

Leave a comment

Your email address will not be published.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