Deprecate use_label_encoder in XGBClassifier. (#7822)

* Deprecate `use_label_encoder` in XGBClassifier.

* We have removed the encoder, now prepare to remove the indicator.

demo/guide-python/cat_in_the_dat.py

@@ -63,7 +63,6 @@ def load_cat_in_the_dat() -> tuple[pd.DataFrame, pd.Series]:
 
 params = {
     "tree_method": "gpu_hist",
-    "use_label_encoder": False,
     "n_estimators": 32,
     "colsample_bylevel": 0.7,
 }

demo/guide-python/continuation.py

@@ -14,13 +14,13 @@ def training_continuation(tmpdir: str, use_pickle: bool) -> None:
     """Basic training continuation."""
     # Train 128 iterations in 1 session
     X, y = load_breast_cancer(return_X_y=True)
-    clf = xgboost.XGBClassifier(n_estimators=128, use_label_encoder=False)
+    clf = xgboost.XGBClassifier(n_estimators=128)
     clf.fit(X, y, eval_set=[(X, y)], eval_metric="logloss")
     print("Total boosted rounds:", clf.get_booster().num_boosted_rounds())
 
     # Train 128 iterations in 2 sessions, with the first one runs for 32 iterations and
     # the second one runs for 96 iterations
-    clf = xgboost.XGBClassifier(n_estimators=32, use_label_encoder=False)
+    clf = xgboost.XGBClassifier(n_estimators=32)
     clf.fit(X, y, eval_set=[(X, y)], eval_metric="logloss")
     assert clf.get_booster().num_boosted_rounds() == 32
 
@@ -54,14 +54,14 @@ def training_continuation_early_stop(tmpdir: str, use_pickle: bool) -> None:
     n_estimators = 512
 
     X, y = load_breast_cancer(return_X_y=True)
-    clf = xgboost.XGBClassifier(n_estimators=n_estimators, use_label_encoder=False)
+    clf = xgboost.XGBClassifier(n_estimators=n_estimators)
     clf.fit(X, y, eval_set=[(X, y)], eval_metric="logloss", callbacks=[early_stop])
     print("Total boosted rounds:", clf.get_booster().num_boosted_rounds())
     best = clf.best_iteration
 
     # Train 512 iterations in 2 sessions, with the first one runs for 128 iterations and
     # the second one runs until early stop.
-    clf = xgboost.XGBClassifier(n_estimators=128, use_label_encoder=False)
+    clf = xgboost.XGBClassifier(n_estimators=128)
     # Reinitialize the early stop callback
     early_stop = xgboost.callback.EarlyStopping(
         rounds=early_stopping_rounds, save_best=True
@@ -79,15 +79,13 @@ def training_continuation_early_stop(tmpdir: str, use_pickle: bool) -> None:
     else:
         path = os.path.join(tmpdir, "model-first-128.json")
         clf.save_model(path)
-        loaded = xgboost.XGBClassifier(use_label_encoder=False)
+        loaded = xgboost.XGBClassifier()
         loaded.load_model(path)
 
     early_stop = xgboost.callback.EarlyStopping(
         rounds=early_stopping_rounds, save_best=True
     )
-    clf = xgboost.XGBClassifier(
+    clf = xgboost.XGBClassifier(n_estimators=n_estimators - 128)
-        n_estimators=n_estimators - 128, use_label_encoder=False
-    )
     clf.fit(
         X,
         y,

demo/guide-python/predict_first_ntree.py

@@ -35,7 +35,7 @@ def native_interface():
 def sklearn_interface():
     X_train, y_train = load_svmlight_file(train)
     X_test, y_test = load_svmlight_file(test)
-    clf = xgb.XGBClassifier(n_estimators=3, max_depth=2, eta=1, use_label_encoder=False)
+    clf = xgb.XGBClassifier(n_estimators=3, max_depth=2, eta=1)
     clf.fit(X_train, y_train, eval_set=[(X_test, y_test)])
     assert clf.n_classes_ == 2
 

doc/tutorials/categorical.rst

@@ -36,9 +36,7 @@ parameter ``enable_categorical``:
 .. code:: python
 
   # Supported tree methods are `gpu_hist`, `approx`, and `hist`.
-  clf = xgb.XGBClassifier(
+  clf = xgb.XGBClassifier(tree_method="gpu_hist", enable_categorical=True)
-      tree_method="gpu_hist", enable_categorical=True, use_label_encoder=False
-  )
   # X is the dataframe we created in previous snippet
   clf.fit(X, y)
   # Must use JSON/UBJSON for serialization, otherwise the information is lost.

python-package/xgboost/sklearn.py

@@ -1304,13 +1304,15 @@ class XGBClassifier(XGBModel, XGBClassifierBase):
         self,
         *,
         objective: _SklObjective = "binary:logistic",
-        use_label_encoder: bool = False,
+        use_label_encoder: Optional[bool] = None,
         **kwargs: Any
     ) -> None:
         # must match the parameters for `get_params`
         self.use_label_encoder = use_label_encoder
         if use_label_encoder is True:
             raise ValueError("Label encoder was removed in 1.6.")
+        if use_label_encoder is not None:
+            warnings.warn("`use_label_encoder` is deprecated in 2.0.0.")
         super().__init__(objective=objective, **kwargs)
 
     @_deprecate_positional_args

tests/python/test_training_continuation.py

@@ -152,16 +152,16 @@ class TestTrainingContinuation:
     def test_changed_parameter(self):
         from sklearn.datasets import load_breast_cancer
         X, y = load_breast_cancer(return_X_y=True)
-        clf = xgb.XGBClassifier(n_estimators=2, use_label_encoder=False)
+        clf = xgb.XGBClassifier(n_estimators=2)
         clf.fit(X, y, eval_set=[(X, y)], eval_metric="logloss")
         assert tm.non_increasing(clf.evals_result()["validation_0"]["logloss"])
 
         with tempfile.TemporaryDirectory() as tmpdir:
             clf.save_model(os.path.join(tmpdir, "clf.json"))
-            loaded = xgb.XGBClassifier(use_label_encoder=False)
+            loaded = xgb.XGBClassifier()
             loaded.load_model(os.path.join(tmpdir, "clf.json"))
 
-        clf = xgb.XGBClassifier(n_estimators=2, use_label_encoder=False)
+        clf = xgb.XGBClassifier(n_estimators=2)
         # change metric to error
         clf.fit(X, y, eval_set=[(X, y)], eval_metric="error")
         assert tm.non_increasing(clf.evals_result()["validation_0"]["error"])

tests/python/test_with_dask.py

@@ -777,9 +777,7 @@ def run_empty_dmatrix_auc(client: "Client", tree_method: str, n_workers: int) ->
     valid_X = dd.from_array(valid_X_, chunksize=n_samples)
     valid_y = dd.from_array(valid_y_, chunksize=n_samples)
 
-    cls = xgb.dask.DaskXGBClassifier(
+    cls = xgb.dask.DaskXGBClassifier(tree_method=tree_method, n_estimators=2)
-        tree_method=tree_method, n_estimators=2, use_label_encoder=False
-    )
     cls.fit(X, y, eval_metric=["auc", "aucpr"], eval_set=[(valid_X, valid_y)])
 
     # multiclass
@@ -808,9 +806,7 @@ def run_empty_dmatrix_auc(client: "Client", tree_method: str, n_workers: int) ->
     valid_X = dd.from_array(valid_X_, chunksize=n_samples)
     valid_y = dd.from_array(valid_y_, chunksize=n_samples)
 
-    cls = xgb.dask.DaskXGBClassifier(
+    cls = xgb.dask.DaskXGBClassifier(tree_method=tree_method, n_estimators=2)
-        tree_method=tree_method, n_estimators=2, use_label_encoder=False
-    )
     cls.fit(X, y, eval_metric=["auc", "aucpr"], eval_set=[(valid_X, valid_y)])
 
 
@@ -837,14 +833,10 @@ def run_auc(client: "Client", tree_method: str) -> None:
     valid_X = dd.from_array(valid_X_, chunksize=10)
     valid_y = dd.from_array(valid_y_, chunksize=10)
 
-    cls = xgb.XGBClassifier(
+    cls = xgb.XGBClassifier(tree_method=tree_method, n_estimators=2)
-        tree_method=tree_method, n_estimators=2, use_label_encoder=False
-    )
     cls.fit(X_, y_, eval_metric="auc", eval_set=[(valid_X_, valid_y_)])
 
-    dcls = xgb.dask.DaskXGBClassifier(
+    dcls = xgb.dask.DaskXGBClassifier(tree_method=tree_method, n_estimators=2)
-        tree_method=tree_method, n_estimators=2, use_label_encoder=False
-    )
     dcls.fit(X, y, eval_metric="auc", eval_set=[(valid_X, valid_y)])
 
     approx = dcls.evals_result()["validation_0"]["auc"]
@@ -1693,7 +1685,6 @@ def test_parallel_submits(client: "Client") -> None:
             verbosity=1,
             n_estimators=i + 1,
             eval_metric="merror",
-            use_label_encoder=False,
         )
         f = client.submit(cls.fit, X, y, pure=False)
         futures.append(f)
@@ -1786,7 +1777,6 @@ def test_parallel_submit_multi_clients() -> None:
                 verbosity=1,
                 n_estimators=i + 1,
                 eval_metric="merror",
-                use_label_encoder=False,
             )
             f = client.submit(cls.fit, X, y, pure=False)
             futures.append((client, f))

tests/python/test_with_sklearn.py

@@ -774,13 +774,12 @@ def save_load_model(model_path):
     X = digits['data']
     kf = KFold(n_splits=2, shuffle=True, random_state=rng)
     for train_index, test_index in kf.split(X, y):
-        xgb_model = xgb.XGBClassifier(use_label_encoder=False).fit(X[train_index], y[train_index])
+        xgb_model = xgb.XGBClassifier().fit(X[train_index], y[train_index])
         xgb_model.save_model(model_path)
 
         xgb_model = xgb.XGBClassifier()
         xgb_model.load_model(model_path)
 
-        assert xgb_model.use_label_encoder is False
         assert isinstance(xgb_model.classes_, np.ndarray)
         assert isinstance(xgb_model._Booster, xgb.Booster)
 
@@ -972,8 +971,8 @@ def test_deprecate_position_arg():
         model.fit(X, y, w)
 
     with pytest.warns(FutureWarning):
-        xgb.XGBClassifier(1, use_label_encoder=False)
+        xgb.XGBClassifier(1)
-    model = xgb.XGBClassifier(n_estimators=1, use_label_encoder=False)
+    model = xgb.XGBClassifier(n_estimators=1)
     with pytest.warns(FutureWarning):
         model.fit(X, y, w)
 
@@ -990,9 +989,6 @@ def test_deprecate_position_arg():
     with pytest.warns(FutureWarning):
         model.fit(X, y, w)
 
-    with pytest.raises(ValueError):
-        xgb.XGBRFClassifier(1, use_label_encoder=True)
-
     model = xgb.XGBRFClassifier(n_estimators=1)
     with pytest.warns(FutureWarning):
         model.fit(X, y, w)
@@ -1334,7 +1330,6 @@ def test_evaluation_metric():
     X, y = load_digits(n_class=10, return_X_y=True)
 
     clf = xgb.XGBClassifier(
-        use_label_encoder=False,
         tree_method="hist",
         eval_metric=merror,
         n_estimators=16,
@@ -1344,7 +1339,6 @@ def test_evaluation_metric():
     custom = clf.evals_result()
 
     clf = xgb.XGBClassifier(
-        use_label_encoder=False,
         tree_method="hist",
         eval_metric="merror",
         n_estimators=16,
@@ -1360,7 +1354,6 @@ def test_evaluation_metric():
     )
 
     clf = xgb.XGBRFClassifier(
-        use_label_encoder=False,
         tree_method="hist", n_estimators=16,
         objective=tm.softprob_obj(10),
         eval_metric=merror,