摘要： 在介绍数据挖掘、分类算法的有关概念的基础上，描述了一个基于SLIQ(Supervised Learning in Quest)的分类算法的实现。算法控制结构采用队列来控制广度优先搜索，对属性表进行内部排序生成有序的属性表序列，再对所有属性表进行遍历，找出所有叶子节点的最佳分裂方案，对应该分裂的类表进行更改，最后生成的决策树通过前序遍历的方式存入输出表。

关键词: 数据挖掘, 分类, 决策树, SLIQ

Abstract: This article introduces the conception of classification arithmetic, describes a kind of implementation of it based on SLIQ. Queue is used in the algorithm control structure to control the breadth-first search. And then, ordered attribute table sequence is formed through the internal sorting of the attribute table. Following, the best split project of the leaf node is found through the traversal of the entire attribute table, and the table, which ought to be splited, is changed. At last, decision-making tree is stored in output table by preorder traversal.

中图分类号: 

• TP311.131