objectif

Régression polynomiale régularisée via pipeline.

code minimal

from sklearn.preprocessing import PolynomialFeatures, StandardScaler
from sklearn.pipeline import Pipeline
from sklearn.linear_model import Ridge
import numpy as np

X = np.arange(10).reshape(-1,1).astype(float); y = X.ravel()**2 + 1
pipe = Pipeline([("poly", PolynomialFeatures(degree=2)), ("sc", StandardScaler(with_mean=False)), ("mdl", Ridge(alpha=1.0))]).fit(X, y)
print(pipe.predict([[3.0]])[0] >= 0.0)

utilisation

print(isinstance(pipe.named_steps["poly"], PolynomialFeatures))

variante(s) utile(s)

from sklearn.linear_model import Lasso
print(hasattr(Lasso(alpha=0.1), "fit"))

notes

• with_mean=False quand sparse; ridge souvent plus stable.