🤖 ML Models
ML - Regularization
Techniques that constrain model complexity to reduce overfitting. The mathematical implementation of "keep it simple." Without regularization, models are free to fit arbitrarily complex patterns — including noise.
2
Minutes
8
Concepts
+45
XP
1
L1 Regularization (Lasso)
Adds the sum of absolute weight values to the loss function:
Loss = Original Loss + λ × Σ|weights|
Effect: Drives some weights to exactly zero — automatic feature selection.
Use when: You suspect many features are irrelevant and want a sparse model.
Typical λ: 1e-4 to 1e-1 (tune via cross-validation).