Building Decision Trees From Decision Stumps
Decision stumps are simple one layer decision trees using all the same settings as the typical decision trees. However, most decision stumps are not useful in prediction on their own. To produce more predictive models, a decision tree is better. Building a decision tree in SAS is much more difficult than a decision stump. By utilizing an innovative approach on macro usage, a full decision tree can be built using a variety of decision stumps. This paper will demonstrate the simplicity of use of the decision stumps.