FIX: ensure that feature importances are properly scaled

Author: glouppe

sklearn/ensemble/forest.py

@@ -335,7 +335,8 @@ def feature_importances_(self):
         all_importances = Parallel(n_jobs=self.n_jobs, backend="threading")(
             delayed(getattr)(tree, 'feature_importances_')
             for tree in self.estimators_)
-        return sum(all_importances) / self.n_estimators
+
+        return sum(all_importances) / len(self.estimators_)
 
 
 class ForestClassifier(six.with_metaclass(ABCMeta, BaseForest,