In the latest version of the document(http://scikit-learn.org/dev/modules/generated/sklearn.tree.DecisionTreeClassifier.html#sklearn.tree.DecisionTreeClassifier), there isn't any explanation of parameter min_impurity_split in DecisionTreeRegressor and DecisionTreeClassifier, which still exists in the current version. In the previous version(http://scikit-learn.org/stable/modules/generated/sklearn.tree.DecisionTreeClassifier.html#sklearn.tree.DecisionTreeClassifier), there are relevant explanations of the parameter. Should we explain it in the latest version like the previous version? If so, I'm willing to help. Thanks.