reflect scikit-learn#17176: zero weight in SV models means that a sample is never a support vector

ivannz · ivannz · commit 1da7c0e245bf · 2020-08-05T20:42:31.000+03:00
diff --git a/sklearn/svm/_classes.py b/sklearn/svm/_classes.py
@@ -1673,3 +1673,11 @@ def predict(self, X):
         """
         y = super().predict(X)
         return np.asarray(y, dtype=np.intp)
+
+    def _more_tags(self):
+        return {
+            '_xfail_checks': {
+                'check_sample_weights_invariance(kind=zeros)':
+                'zero sample_weight is not equivalent to removing samples',
+            }
+        }
