Do you need to standardize WoE-encoded variables when using L1 regularized Logistic Regression?

91 Views Asked by At

I read that numerical variables should be standardized before a Logistic Regression with L1 Lasso Regularization is trained on them. If you have a categorical variable, it is recommended to encode it with Weight of Evidence. This transforms the former categorical variable to a numeric one. Should you standardize it then too?

0

There are 0 best solutions below