Native TargetEncoder: target encoding without leakage
scikit-learn's TargetEncoder applies internal cross-fitting during fit_transform: each row is encoded using means computed without it, which neutralizes target leakage.
Prerequisites
scikit-learn 1.3+
Python
from sklearn.preprocessing import TargetEncoder
from sklearn.pipeline import Pipeline
from sklearn.ensemble import HistGradientBoostingClassifier
from sklearn.compose import ColumnTransformer
prep = ColumnTransformer([
("target_enc", TargetEncoder(cv=5, smooth="auto"), ["ville", "code_naf"]),
], remainder="passthrough")
pipe = Pipeline([
("prep", prep),
("clf", HistGradientBoostingClassifier(random_state=42)),
])
# fit_transform utilise le cross-fitting ; transform (en test) utilise
# l'encodage appris sur tout le train -> aucune fuite.
pipe.fit(X_train, y_train)Result
Pipeline(steps=[('prep',
ColumnTransformer(remainder='passthrough',
transformers=[('target_enc',
TargetEncoder(cv=5,
smooth='auto'),
['ville', 'code_naf'])])),
('clf', HistGradientBoostingClassifier(random_state=42))])TargetEncoderFuite de donnéesCardinalitéEncodage