Which regularization method tends to shrink coefficients toward zero without necessarily setting them to zero?

Which regularization method tends to shrink coefficients toward zero without necessarily setting them to zero?

• A. LASSO
• B. Ridge
• C. Multicollinearity
• D. Outliers

Answer: (B)

Regularization adds a penalty to the loss to keep coefficients small. Ridge regression uses an L2 penalty, which squares the coefficients and shrinks them toward zero. This tends to reduce the size of all coefficients, especially for less informative features, but it usually doesn’t drive them exactly to zero, so all predictors stay in the model with smaller weights. This helps reduce overfitting and stabilizes estimates, particularly when predictors are correlated. In contrast, LASSO uses an L1 penalty and can push some coefficients all the way to zero, effectively removing those features. Multicollinearity and outliers are data characteristics, not regularization methods.

Regularization adds a penalty to the loss to keep coefficients small. Ridge regression uses an L2 penalty, which squares the coefficients and shrinks them toward zero. This tends to reduce the size of all coefficients, especially for less informative features, but it usually doesn’t drive them exactly to zero, so all predictors stay in the model with smaller weights. This helps reduce overfitting and stabilizes estimates, particularly when predictors are correlated. In contrast, LASSO uses an L1 penalty and can push some coefficients all the way to zero, effectively removing those features. Multicollinearity and outliers are data characteristics, not regularization methods.