8000 [MRG] Set min_impurity_split in gradient boosting models (#8007) · scikit-learn/scikit-learn@3162f98 · GitHub
[go: up one dir, main page]
Skip to content

Commit 3162f98

Browse files
sebpjnothman
sebp
authored and
jnothman
committed
[MRG] Set min_impurity_split in gradient boosting models (#8007)
 Fixes #8006
1 parent aceb73c commit 3162f98

File tree

3 files changed

+19
-1
lines changed

3 files changed

+19
-1
lines changed

‎doc/whats_new.rst

Lines changed: 5 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -118,6 +118,11 @@ Bug fixes
118118
when a numpy array is passed in for weights. :issue:`7983` by
119119
:user:`Vincent Pham <vincentpham1991>`.
120120

121+
- Fix a bug where :class:`sklearn.ensemble.GradientBoostingClassifier` and
122+
:class:`sklearn.ensemble.GradientBoostingRegressor` ignored the
123+
``min_impurity_split`` parameter.
124+
:issue:`8006` by :user:`Sebastian Pölsterl <sebp>`.
125+
121126
API changes summary
122127
-------------------
123128

@@ -127,7 +132,6 @@ API changes summary
127132
now only have ``self.estimators_`` available after ``fit``.
128133
:issue:`7464` by `Lars Buitinck`_ and `Loic Esteve`_.
129134

130-
131135
.. _changes_0_18_1:
132136

133137
Version 0.18.1

‎sklearn/ensemble/gradient_boosting.py

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -765,6 +765,7 @@ def _fit_stage(self, i, X, y, y_pred, sample_weight, sample_mask,
765765
min_samples_split=self.min_samples_split,
766766
min_samples_leaf=self.min_samples_leaf,
767767
min_weight_fraction_leaf=self.min_weight_fraction_leaf,
768+
min_impurity_split=self.min_impurity_split,
768769
max_features=self.max_features,
769770
max_leaf_nodes=self.max_leaf_nodes,
770771
random_state=random_state,

‎sklearn/ensemble/tests/test_gradient_boosting.py

Lines changed: 13 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -961,6 +961,19 @@ def test_max_leaf_nodes_max_depth():
961961
assert_equal(tree.max_depth, 1)
962962

963963

964+
def test_min_impurity_split():
965+
# Test if min_impurity_split of base estimators is set
966+
# Regression test for #8006
967+
X, y = datasets.make_hastie_10_2(n_samples=100, random_state=1)
968+
all_estimators = [GradientBoostingRegressor,
969+
GradientBoostingClassifier]
970+
971+
for GBEstimator in all_estimators:
972+
est = GBEstimator(min_impurity_split=0.1).fit(X, y)
973+
for tree in est.estimators_.flat:
974+
assert_equal(tree.min_impurity_split, 0.1)
975+
976+
964977
def test_warm_start_wo_nestimators_change():
965978
# Test if warm_start does nothing if n_estimators is not changed.
966979
# Regression test for #3513.

0 commit comments

Comments
 (0)
0