摘要： 通过分析对象属性的关系,在建立属性列表的基础上,简化有效属性,同时利用分组计数的方法统计属性取值的类别分布信息,提出了一个基于决策树的两阶段多属性分类算法,可以有效地提高发现分类规则的准确性。最后给出了相应的具体算法。

关键词: 数据挖掘, 分类分析, 多属性, 决策树

Abstract: The paper analyses the relation of object attributes to simplify the valid attributes in dataset based on attribute list, and gains the class information by the method of grouping count. A algorithm on multi-attribute classifying based on tree is put forward to improve the veracity of classifying. Finally, corresponding algorithm is developed.

中图分类号: 

• TP311