L2 Regularization (Ridge)
← Back to Linear Models
Adds the sum of squared weights as a penalty term. Shrinks all weights toward zero but does not eliminate any. Handles multicollinearity well and generally produces more stable models than unregularized alternatives.