Add TargetEncoder for scikit-learn

Author: x99b

stubs/sklearn/preprocessing/__init__.pyi

@@ -19,7 +19,7 @@ from ._data import (
     scale as scale,
 )
 from ._discretization import KBinsDiscretizer as KBinsDiscretizer
-from ._encoders import OneHotEncoder as OneHotEncoder, OrdinalEncoder as OrdinalEncoder
+from ._encoders import OneHotEncoder as OneHotEncoder, OrdinalEncoder as OrdinalEncoder, TargetEncoder as TargetEncoder
 from ._function_transformer import FunctionTransformer as FunctionTransformer
 from ._label import (
     LabelBinarizer as LabelBinarizer,
@@ -47,6 +47,7 @@ __all__ = [
     "RobustScaler",
     "SplineTransformer",
     "StandardScaler",
+    "TargetEncoder",
     "add_dummy_feature",
     "PolynomialFeatures",
     "binarize",

stubs/sklearn/preprocessing/_encoders.pyi

@@ -13,7 +13,7 @@ from ..base import BaseEstimator, OneToOneFeatureMixin, TransformerMixin
 #          Joris Van den Bossche <jorisvandenbossche@gmail.com>
 # License: BSD 3 clause
 
-__all__ = ["OneHotEncoder", "OrdinalEncoder"]
+__all__ = ["OneHotEncoder", "OrdinalEncoder", "TargetEncoder"]
 
 class _BaseEncoder(TransformerMixin, BaseEstimator): ...
 
@@ -67,3 +67,29 @@ class OrdinalEncoder(OneToOneFeatureMixin, _BaseEncoder):
     def fit(self, X: MatrixLike, y: Series | None = None) -> Self: ...
     def transform(self, X: MatrixLike) -> ndarray: ...
     def inverse_transform(self, X: MatrixLike) -> ndarray: ...
+
+class TargetEncoder(OneToOneFeatureMixin, _BaseEncoder):
+    feature_names_in_: ndarray = ...
+    n_features_in_: int = ...
+    categories_: list[ndarray] = ...
+    encodings_: list[ndarray] = ...
+    target_type_: str = ...
+    target_mean_: float = ...
+    classes_: ndarray | None = ...
+
+    _parameter_constraints: ClassVar[dict] = ...
+
+    def __init__(
+        self,
+        *,
+        categories: Sequence[ArrayLike] | Literal["auto"] = "auto",
+        target_type: Literal["auto", "continuous", "binary", "multiclass"] = "auto",
+        smooth: Literal["auto"] | float = "auto",
+        cv: int = 5,
+        shuffle: bool = True,
+        random_state: Int | None = None,
+    ) -> None: ...
+    def fit(self, X: MatrixLike, y: ArrayLike) -> Self: ...
+    def transform(self, X: MatrixLike) -> ndarray: ...
+    def fit_transform(self, X: MatrixLike, y: ArrayLike) -> ndarray: ...
+    def get_feature_names_out(self, input_features: None | ArrayLike = None) -> ndarray: ...