It should be noted that when selecting splits, classification and regression trees may track the competitive splits at each decision point along the way. A competitive split is one that results in nearly as pure a node as the chosen split. Classification and regression trees may also keep track of surrogate variables. Use of a surrogate variable at a given split results in a similar node impurity measure (as would a competitor) but also mimics the chosen split itself in terms of which and how many observations go which way in the split.
Was this article helpful?