最大频繁子图挖掘算法研究  

Research on the Mining Algorithms for Maximal Frequent Subgraphs

作　　者：李继腾[1] 骆志刚[1] 丁凡[1] 田文颖[1] 赵琦[1] 

机构地区：[1]国防科技大学计算机学院,湖南长沙410073

出　　处：《计算机工程与科学》2009年第12期67-70,共4页Computer Engineering & Science

基　　金：国家863计划资助项目(2007AA01Z106);国家自然科学基金资助项目(60673018)

摘　　要：随着图的广泛应用,图的规模不断扩大,因此提高频繁子图挖掘效率势在必行。本文针对频繁子图挖掘所产生的庞大的结果集,提出了一个最大频繁子图挖掘算法MFME,从而极大地减少了结果集的数量。MFME使用了映射的思想将图集中的边映射到边表中并在此表上进行子图挖掘,有效地提高了算法的效率。实验结果表明,MFME的效率较经典算法SPIN有明显提高。With the extensive application of graphs, their sizes are expanding unceasingly. Therefore it is imperative to improve the efficiency of mining the frequent subgraphs. According to the huge number of possible subgraphs, this paper proposes an algorithm MFME for mining maximal frequent subgraphs, which greatly reduces the number of the subgraph sets. The algorithm MFME which is based on the idea of mapping focuses on mapping the edge from the graph set to the edge table and it improves the efficiency of the algorithm effectively. The experimental results show that MFME is more efficient than algorithm SPIN.

关 键 词：数据挖掘 频繁子图 子图同构 映射树 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]