Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[skl] Enable cat feature without specifying tree method. (#9353)

Browse Source
This commit is contained in:
Jiaming Yuan 2023-07-03 22:06:17 +08:00 committed by GitHub
parent 39390cc2ee
commit e964654b8f
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 5 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
python-package/xgboost/sklearn.py
View File

@ -930,8 +930,7 @@ class XGBModel(XGBModelBase):
callbacks = self.callbacks if self.callbacks is not None else callbacks callbacks = self.callbacks if self.callbacks is not None else callbacks
tree_method = params.get("tree_method", None) tree_method = params.get("tree_method", None)
cat_support = {"gpu_hist", "approx", "hist"} if self.enable_categorical and tree_method == "exact":
if self.enable_categorical and tree_method not in cat_support:
raise ValueError( raise ValueError(
"Experimental support for categorical data is not implemented for" "Experimental support for categorical data is not implemented for"
" current tree method yet." " current tree method yet."

3
tests/python/test_with_sklearn.py
View File

@ -1390,7 +1390,6 @@ def test_categorical():
X, y = tm.make_categorical(n_samples=32, n_features=2, n_categories=3, onehot=False) X, y = tm.make_categorical(n_samples=32, n_features=2, n_categories=3, onehot=False)
ft = ["c"] * X.shape[1] ft = ["c"] * X.shape[1]
reg = xgb.XGBRegressor( reg = xgb.XGBRegressor(
tree_method="hist",
feature_types=ft, feature_types=ft,
max_cat_to_onehot=1, max_cat_to_onehot=1,
enable_categorical=True, enable_categorical=True,
@ -1409,7 +1408,7 @@ def test_categorical():
onehot, y = tm.make_categorical( onehot, y = tm.make_categorical(
n_samples=32, n_features=2, n_categories=3, onehot=True n_samples=32, n_features=2, n_categories=3, onehot=True
) )
reg = xgb.XGBRegressor(tree_method="hist") reg = xgb.XGBRegressor()
reg.fit(onehot, y, eval_set=[(onehot, y)]) reg.fit(onehot, y, eval_set=[(onehot, y)])
from_enc = reg.evals_result()["validation_0"]["rmse"] from_enc = reg.evals_result()["validation_0"]["rmse"]
predt_enc = reg.predict(onehot) predt_enc = reg.predict(onehot)

6
tests/test_distributed/test_with_dask/test_with_dask.py
View File

@ -308,7 +308,7 @@ def test_dask_sparse(client: "Client") -> None:
def run_categorical(client: "Client", tree_method: str, X, X_onehot, y) -> None: def run_categorical(client: "Client", tree_method: str, X, X_onehot, y) -> None:
parameters = {"tree_method": tree_method, "max_cat_to_onehot": 9999} # force onehot parameters = {"tree_method": tree_method, "max_cat_to_onehot": 9999} # force onehot
rounds = 10 rounds = 10
m = xgb.dask.DaskDMatrix(client, X_onehot, y, enable_categorical=True) m = xgb.dask.DaskDMatrix(client, X_onehot, y, enable_categorical=True)
by_etl_results = xgb.dask.train( by_etl_results = xgb.dask.train(
@ -364,9 +364,9 @@ def run_categorical(client: "Client", tree_method: str, X, X_onehot, y) -> None:
check_model_output(reg.get_booster()) check_model_output(reg.get_booster())
reg = xgb.dask.DaskXGBRegressor( reg = xgb.dask.DaskXGBRegressor(
enable_categorical=True, n_estimators=10 enable_categorical=True, n_estimators=10, tree_method="exact"
) )
with pytest.raises(ValueError): with pytest.raises(ValueError, match="categorical data"):
reg.fit(X, y) reg.fit(X, y)
# check partition based # check partition based
reg = xgb.dask.DaskXGBRegressor( reg = xgb.dask.DaskXGBRegressor(