Skip to main content

👮 Regularization

Regularization is the "Police Force" of ML. It stops the model from cheating!

⚖️ The Penalty​

It forces the model to find simple, smooth rules instead of jagged memorization lines (overfitting).

  • L1 (Lasso): Forces useless weights to zero (deletes useless features).
  • L2 (Ridge): Shrinks all weights to be very small.

🐍 Python Implementation​

from sklearn.linear_model import Ridge # L2 Regularization

# Alpha is the "strength" of the police force
model = Ridge(alpha=1.0)
# model.fit(X, y)