一种基于Apriori思想的频繁子图发现算法  

A Frequent Subgraph Discovery Algorithm Based on Apriori’s Idea

作　　者：李玉华[1] 罗汉果[1] 孙小林[1] 

机构地区：[1]华中科技大学计算机科学与技术学院,湖北武汉430074

出　　处：《计算机工程与科学》2007年第4期84-87,共4页Computer Engineering & Science

基　　金：国家自然科学基金资助项目(60403027);国家"十五"重大科技攻关项目(2002BA103A04)

摘　　要：如今,关联规则技术应用在许多非传统领域,许多已有的频繁项集搜索方法已经不适用了。一种解决的方法就是用图的形式表示这些领域的事务,然后利用基于图论的数据挖掘技术发现频繁子图。本文提出了一种基于Aproiri思想的频繁子图发现算法SLAGM,它可以有效地挖掘简单图中的频繁子图。实验证明,该算法在性能上优于另一种子图挖掘算法AGM。Nowadays association rule mining techniques are applied to many non-traditional domains, existing approaches for frequent itemsets discovery are inapplicable. An alternate way to solve these problems is to represent the transactions of those domains by graph,and find the frequent subgraphs by using graph-based data mining techniques. We propose a new algorithm named SLAGM, which is based on Apriori＇s idea. It can mine frequent subgraphs from simple graphs efficiently. After being evaluated by experiments with synthetic datasets, this algorithm show better performance than another subgraph mining algorithm AGM.

关 键 词：图论 频繁子图 数据挖掘 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]