What is the role of regularization in Scikit-Learn linear models?

Regularization penalizes large coefficients to prevent overfitting.

Linear models in Scikit-Learn use L1 (Lasso), L2 (Ridge), or ElasticNet regularization. These penalties are added to the loss function to constrain coefficient magnitude, improving generalization. L1 can produce sparse models, while L2 distributes penalty smoothly.