Which regularization method uses an L2 penalty to shrink coefficients but not drop inputs?

Which regularization method uses an L2 penalty to shrink coefficients but not drop inputs?

• A. LASSO (L1 penalty)
• B. Ridge Regression (L2 penalty)
• C. Elastic Net (L1 + L2 penalties)
• D. K-fold Cross-Validation

Answer: (B)

Ridge regression applies an L2 penalty to the loss function, shrinking coefficients toward zero but not eliminating inputs. This means every feature contributes to the model, just with a smaller weight, which helps reduce variance without creating a sparse model. In contrast, an L1 penalty (LASSO) tends to set some coefficients to exactly zero, effectively dropping those inputs, while Elastic Net mixes both penalties. K-fold cross-validation is a evaluation method for estimating generalization performance and tuning hyperparameters, not a regularization penalty. So the method that uses an L2 penalty to shrink coefficients but keep all inputs is Ridge regression.

Ridge regression applies an L2 penalty to the loss function, shrinking coefficients toward zero but not eliminating inputs. This means every feature contributes to the model, just with a smaller weight, which helps reduce variance without creating a sparse model. In contrast, an L1 penalty (LASSO) tends to set some coefficients to exactly zero, effectively dropping those inputs, while Elastic Net mixes both penalties. K-fold cross-validation is a evaluation method for estimating generalization performance and tuning hyperparameters, not a regularization penalty. So the method that uses an L2 penalty to shrink coefficients but keep all inputs is Ridge regression.