首页 | 本学科首页   官方微博 | 高级检索  
     

基于属性间交互信息的ID3算法
引用本文:谢竞博,王熙照.基于属性间交互信息的ID3算法[J].计算机工程与应用,2004,40(30):93-94,150.
作者姓名:谢竞博  王熙照
作者单位:河北大学数学与计算机学院,保定,071002
基金项目:河北省自然科学基金资助项目(编号:603137)
摘    要:启发式算法是决策树研究的核心。文中分析了最常见的一种决策树归纳启发式算法即ID3算法的不足,给出了一个改进版本,它在选择测试属性时不仅要求该属性带来的信息增益尽可能大,而且要求其与同一分支上已经使用过的各属性之间的交互信息尽可能小,从而避免了对冗余属性的选择,实现信息熵的真正减少。分析及实验结果表明,与ID3算法相比,该算法能构造出更优的决策树。

关 键 词:决策树  ID3  启发式算法  交互信息
文章编号:1002-8331-(2004)30-0093-02

An Extended Heuristic Algorithm to ID3 Based on the Mutual Information between Attributes
Xie Jingbo,Wang Xizhao.An Extended Heuristic Algorithm to ID3 Based on the Mutual Information between Attributes[J].Computer Engineering and Applications,2004,40(30):93-94,150.
Authors:Xie Jingbo  Wang Xizhao
Abstract:Heuristic algorithm is the kernel in the research on decision tree.The paper analyzes the shortcomings of the most popular heuristic algorithm for generating decision tree,ie.ID3,and gives an extended version to it,which demandsthat the selected attribute can bring more information gain as well as less mutual information between it with others on its ancestor nodes.The algorithm can avoid selecting the redundant attributes and achieve the real reduce in entropy.Analysis and experiments indicate:compared with ID3,the extended heuristic algorithm can construct a better tree.
Keywords:decision tree  ID3  heuristic algorithm  mutual information  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号