Decision tree pre-pruning is an effective method to solve the over-fitting problem in decision tree learning process. However, it is difficult to estimate the exact time to stop the growing process of a decision tree, which limits the developments and applications of this method. In this paper, the growing of a decision tree is controlled by the uncertainty of a decision table, and a data-driven learning algorithm for decision tree pre-pruning is developed.
Published in:
Active Media Technology, 2005. (AMT 2005). Proceedings of the 2005 International Conference on
Date of Conference: 19-21 May 2005