scikit-learn · adrinjalali · Jul 31, 2024 · Jun 14, 2024 · Jun 14, 2024 · Jun 14, 2024
diff --git a/doc/metadata_routing.rst b/doc/metadata_routing.rst
@@ -285,6 +285,7 @@ Meta-estimators and functions supporting metadata routing:
 - :class:`sklearn.ensemble.BaggingClassifier`
 - :class:`sklearn.ensemble.BaggingRegressor`
 - :class:`sklearn.feature_selection.SelectFromModel`
+- :class:`sklearn.feature_selection.SequentialFeatureSelector`
 - :class:`sklearn.impute.IterativeImputer`
 - :class:`sklearn.linear_model.ElasticNetCV`
 - :class:`sklearn.linear_model.LarsCV`
@@ -324,4 +325,3 @@ Meta-estimators and tools not supporting metadata routing yet:
 - :class:`sklearn.ensemble.AdaBoostRegressor`
 - :class:`sklearn.feature_selection.RFE`
 - :class:`sklearn.feature_selection.RFECV`
-- :class:`sklearn.feature_selection.SequentialFeatureSelector`
diff --git a/doc/whats_new/v1.6.rst b/doc/whats_new/v1.6.rst
@@ -83,6 +83,11 @@ more details.
   params to the underlying regressor.
   :pr:`29136` by :user:`Omar Salman <OmarManzoor>`.
 
+- |Feature| :class:`feature_selection.SequentialFeatureSelector` now supports
+  metadata routing in its `fit` method and passes the corresponding params to
+  the :func:`model_selection.cross_val_score` function.
+  :pr:`29260` by :user:`Omar Salman <OmarManzoor>`.
+
 - |Feature| :func:`model_selection.validation_curve` now supports metadata routing for
   the `fit` method of its estimator and for its underlying CV splitter and scorer.
   :pr:`29329` by :user:`Stefanie Senger <StefanieSenger>`.

diff --git a/sklearn/feature_selection/_sequential.py b/sklearn/feature_selection/_sequential.py
@@ -10,18 +10,22 @@
 import numpy as np
 
 from ..base import BaseEstimator, MetaEstimatorMixin, _fit_context, clone, is_classifier
-from ..metrics import get_scorer_names
+from ..metrics import check_scoring, get_scorer_names
 from ..model_selection import check_cv, cross_val_score
+from ..utils._metadata_requests import (
+    MetadataRouter,
+    MethodMapping,
+    _raise_for_params,
+    _routing_enabled,
+    process_routing,
+)
 from ..utils._param_validation import HasMethods, Interval, RealNotInt, StrOptions
 from ..utils._tags import _safe_tags
-from ..utils.metadata_routing import _RoutingNotSupportedMixin
 from ..utils.validation import check_is_fitted
 from ._base import SelectorMixin
 
 
-class SequentialFeatureSelector(
-    _RoutingNotSupportedMixin, SelectorMixin, MetaEstimatorMixin, BaseEstimator
-):
+class SequentialFeatureSelector(SelectorMixin, MetaEstimatorMixin, BaseEstimator):
     """Transformer that performs Sequential Feature Selection.
 
     This Sequential Feature Selector adds (forward selection) or
@@ -191,7 +195,7 @@ def __init__(
         # SequentialFeatureSelector.estimator is not validated yet
         prefer_skip_nested_validation=False
     )
-    def fit(self, X, y=None):
+    def fit(self, X, y=None, **params):
         """Learn the features to select from X.
 
         Parameters
@@ -204,11 +208,24 @@ def fit(self, X, y=None):
             Target values. This parameter may be ignored for
             unsupervised learning.
 
+        **params : dict, default=None
+            Parameters to be passed to the underlying `estimator`, `cv`
+            and `scorer` objects.
+
+            .. versionadded:: 1.6
+
+                Only available if `enable_metadata_routing=True`,
+                which can be set by using
+                ``sklearn.set_config(enable_metadata_routing=True)``.
+                See :ref:`Metadata Routing User Guide <metadata_routing>` for
+                more details.
+
         Returns
         -------
         self : object
             Returns the instance itself.
         """
+        _raise_for_params(params, self, "fit")
         tags = self._get_tags()
         X = self._validate_data(
             X,
@@ -251,9 +268,15 @@ def fit(self, X, y=None):
 
         old_score = -np.inf
         is_auto_select = self.tol is not None and self.n_features_to_select == "auto"
+
+        # We only need to verify the routing here and not use the routed params
+        # because internally the actual routing will also take place inside the
+        # `cross_val_score` function.
+        if _routing_enabled():
+            process_routing(self, "fit", **params)
         for _ in range(n_iterations):
             new_feature_idx, new_score = self._get_best_new_feature_score(
-                cloned_estimator, X, y, cv, current_mask
+                cloned_estimator, X, y, cv, current_mask, **params
             )
             if is_auto_select and ((new_score - old_score) < self.tol):
                 break
@@ -269,7 +292,7 @@ def fit(self, X, y=None):
 
         return self
 
-    def _get_best_new_feature_score(self, estimator, X, y, cv, current_mask):
+    def _get_best_new_feature_score(self, estimator, X, y, cv, current_mask, **params):
         # Return the best new feature and its score to add to the current_mask,
         # i.e. return the best new feature and its score to add (resp. remove)
         # when doing forward selection (resp. backward selection).
@@ -290,6 +313,7 @@ def _get_best_new_feature_score(self, estimator, X, y, cv, current_mask):
                 cv=cv,
                 scoring=self.scoring,
                 n_jobs=self.n_jobs,
+                params=params,
             ).mean()
         new_feature_idx = max(scores, key=lambda feature_idx: scores[feature_idx])
         return new_feature_idx, scores[new_feature_idx]
@@ -302,3 +326,32 @@ def _more_tags(self):
         return {
             "allow_nan": _safe_tags(self.estimator, key="allow_nan"),
         }
+
+    def get_metadata_routing(self):
+        """Get metadata routing of this object.
+
+        Please check :ref:`User Guide <metadata_routing>` on how the routing
+        mechanism works.
+
+        .. versionadded:: 1.6
+
+        Returns
+        -------
+        routing : MetadataRouter
+            A :class:`~sklearn.utils.metadata_routing.MetadataRouter` encapsulating
+            routing information.
+        """
+        router = MetadataRouter(owner=self.__class__.__name__)
+        router.add(
+            estimator=self.estimator,
+            method_mapping=MethodMapping().add(caller="fit", callee="fit"),
+        )
+        router.add(
+            splitter=check_cv(self.cv, classifier=is_classifier(self.estimator)),
+            method_mapping=MethodMapping().add(caller="fit", callee="split"),
+        )
+        router.add(
+            scorer=check_scoring(self.estimator, scoring=self.scoring),
+            method_mapping=MethodMapping().add(caller="fit", callee="score"),
+        )
+        return router
diff --git a/sklearn/feature_selection/tests/test_sequential.py b/sklearn/feature_selection/tests/test_sequential.py
@@ -321,3 +321,12 @@ def test_cv_generator_support():
 
     sfs = SequentialFeatureSelector(knc, n_features_to_select=5, cv=splits)
     sfs.fit(X, y)
+
+
+def test_fit_rejects_params_with_no_routing_enabled():
+    X, y = make_classification(random_state=42)
+    est = LinearRegression()
+    sfs = SequentialFeatureSelector(estimator=est)
+
+    with pytest.raises(ValueError, match="is only supported if"):
+        sfs.fit(X, y, sample_weight=np.ones_like(y))
diff --git a/sklearn/tests/test_metaestimators_metadata_routing.py b/sklearn/tests/test_metaestimators_metadata_routing.py
@@ -407,6 +407,18 @@ def enable_slep006():
         ],
         "method_mapping": {"fit": ["fit", "score"]},
     },
+    {
+        "metaestimator": SequentialFeatureSelector,
+        "estimator_name": "estimator",
+        "estimator": "classifier",
+        "X": X,
+        "y": y,
+        "estimator_routing_methods": ["fit"],
+        "scorer_name": "scoring",
+        "scorer_routing_methods": ["fit"],
+        "cv_name": "cv",
+        "cv_routing_methods": ["fit"],
+    },
 ]
 """List containing all metaestimators to be tested and their settings
 
@@ -450,7 +462,6 @@ def enable_slep006():
     AdaBoostRegressor(),
     RFE(ConsumingClassifier()),
     RFECV(ConsumingClassifier()),
-    SequentialFeatureSelector(ConsumingClassifier()),
 ]