计算机应用 ›› 2005, Vol. 25 ›› Issue (09): 1986-1988.DOI: 10.3724/SP.J.1087.2005.01986

• Web与数据库 • 上一篇    下一篇

基于IF-THEN规则的决策树裁剪算法

张晓龙,骆名剑   

  1. 武汉科技大学计算机科学与技术学院
  • 发布日期:2011-04-11 出版日期:2005-09-01
  • 基金资助:

    教育部留学回国人员科研启动基金;;湖北省教育厅重点项目(2004D006)

Decision tree pruning algorithm with IF-THEN rules

ZHANG Xiao-long,LUO Ming-jian   

  1. College of Computer Science & Technology,Wuhan University of Science and Technology,Wuhan Hubei 430081,China
  • Online:2011-04-11 Published:2005-09-01

摘要: 决策树是机器学习和数据挖掘领域中一种基本的学习方法。文中分析了C4.5算法以及该算法不足之处,提出了一种决策树裁剪算法,其中以规则信息量作为判断标准。实验结果表明这种方法可以提高最终模型的预测精度,并能够很好克服数据中的噪音。

关键词: 决策树, 树裁剪, 规则信息量

Abstract: Decision tree is a basic learning method in machine learning and data mining.An algorithm was presented to prune a decision tree with a set of rules that are converted from C4.5,where rule information was used as a pruning criterion.The empirical results demonstrate that the proposed pruning algorithm has high predict accuracy and good noise-tolerance.

Key words: decision tree, tree pruning, rule information

中图分类号: