摘要： 介绍了怎样通过信息量或熵的比较来构造一个决策树的数据挖掘算法，并且就一些特殊的地方进行了讨论分析，例如怎样处理高分枝属性、数值属性和缺失数据以及怎样剪枝．利用模型系统的一些源代码来具体地实现算法中的一些模块，并且描述了国内外的有关数据挖掘的研究情况．

关键词: 数据挖掘, 决策树, 信息量, 熵, 剪枝

Abstract: We introduce the algorithm of how to build a decision tree by the comparison of information value or entropy and how to deal with something special, e.g. how to handle high-branching attributes, numeric attributes, missing values and how to prune. Finally, we show some source code of some modules in our implementation of this algorithm and give some introduction about the research of data mining at home and abroad.

中图分类号: 

• TP311.13