熵在决策树构建中的应用  

The Entropy Theory and Its Application in the Decision Tree

作　　者：杨宝华[1] 

机构地区：[1]安徽农业大学信息与计算机学院,安徽合肥230036

出　　处：《电脑知识与技术》2006年第11期180-181,共2页Computer Knowledge and Technology

基　　金：安徽省自然科学基金项目资助(01041108)

摘　　要：决策树是对未知数据进行分类预测的一种方法。本文阐述熵理论和典型的ID3算法,并举例构建决策树。讨论选择具有最高信息增益的属性构建决策树,即选定具有最高区分度的属性作为当前节点。利用熵理论构建决策树,方法简单,结构清晰,容易转换成分类规则,并且不需要先验领域知识。Decision tree is a kind of predictive-method for data to classify. In this paper the entropy theory and the typical algorithm of the ID3 are elaborated, and an example is proposed to build the decision tree. The attribute with the highest information gain is chosen to make the decision tree, namely the attribute with the highest discrimination as the current node, which is easy to transform the classified rules. The method needn＇t the domain knowledge to be examined and is simple to understand.

关 键 词：熵 决策树 算法 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]