无序嵌入式频繁子树挖掘算法  

Mining Algorithm for Unordered Embedded Frequent Subtree

作　　者：刘波[1] 杨燕[1] 

机构地区：[1]西南交通大学信息科学与技术学院,成都610031

出　　处：《计算机工程》2009年第3期51-53,共3页Computer Engineering

摘　　要：频繁模式挖掘的研究对象包括事务、序列、树和图。该文提出用模式增长方法在无序树构成的森林中挖掘嵌入频繁子树。利用规范化方法实现用唯一的形式表现无序树,根据待增长模式的拓扑结构确定其增长点并构造相应的投影库,将挖掘频繁子树模式问题转化为在各个投影库中寻找频繁节点的问题。Frequent patterns mining involves mining transactions, sequences, trees and graphs. This paper presents an efficient pattern growth algorithm for mining frequent embedded subtrees in a forest composed of unordered trees. It uses a canonical method to represent unordered trees in a unique way. It creates a projection database for every growing point of the pattern to grow. The problem is transformed from mining frequent trees to find frequent nodes in the projection database.

关 键 词：频繁模式 频繁子树 无序树 嵌入式子树 

分 类 号：TP311.12[自动化与计算机技术—计算机软件与理论]