ENH Make GradientBoostingClassifier error message more informative (scikit-learn#10207)

Author: gxyd

sklearn/ensemble/gradient_boosting.py

@@ -1003,7 +1003,7 @@ def fit(self, X, y, sample_weight=None, monitor=None):
 
         check_consistent_length(X, y, sample_weight)
 
-        y = self._validate_y(y)
+        y = self._validate_y(y, sample_weight)
 
         if self.n_iter_no_change is not None:
             X, X_val, y, y_val, sample_weight, sample_weight_val = (
@@ -1237,7 +1237,9 @@ def feature_importances_(self):
         importances = total_sum / len(self.estimators_)
         return importances
 
-    def _validate_y(self, y):
+    def _validate_y(self, y, sample_weight):
+        # 'sample_weight' is not utilised but is used for
+        # consistency with similar method _validate_y of GBC
         self.n_classes_ = 1
         if y.dtype.kind == 'O':
             y = y.astype(np.float64)
@@ -1551,9 +1553,15 @@ def __init__(self, loss='deviance', learning_rate=0.1, n_estimators=100,
             validation_fraction=validation_fraction,
             n_iter_no_change=n_iter_no_change, tol=tol)
 
-    def _validate_y(self, y):
+    def _validate_y(self, y, sample_weight):
         check_classification_targets(y)
         self.classes_, y = np.unique(y, return_inverse=True)
+        n_trim_classes = np.count_nonzero(np.bincount(y, sample_weight))
+        if n_trim_classes < 2:
+            raise ValueError("y contains %d class after sample_weight "
+                             "trimmed classes with zero weights, while a "
+                             "minimum of 2 classes are required."
+                             % n_trim_classes)
         self.n_classes_ = len(self.classes_)
         return y
 

sklearn/ensemble/tests/test_gradient_boosting.py

@@ -26,8 +26,8 @@
 from sklearn.utils.testing import assert_equal
 from sklearn.utils.testing import assert_greater
 from sklearn.utils.testing import assert_less
-from sklearn.utils.testing import assert_raise_message
 from sklearn.utils.testing import assert_raises
+from sklearn.utils.testing import assert_raise_message
 from sklearn.utils.testing import assert_true
 from sklearn.utils.testing import assert_warns
 from sklearn.utils.testing import assert_warns_message
@@ -364,6 +364,12 @@ def test_check_inputs():
     assert_raises(ValueError, clf.fit, X, y,
                   sample_weight=([1] * len(y)) + [0, 1])
 
+    weight = [0, 0, 0, 1, 1, 1]
+    clf = GradientBoostingClassifier(n_estimators=100, random_state=1)
+    msg = ("y contains 1 class after sample_weight trimmed classes with "
+           "zero weights, while a minimum of 2 classes are required.")
+    assert_raise_message(ValueError, msg, clf.fit, X, y, sample_weight=weight)
+
 
 def test_check_inputs_predict():
     # X has wrong shape