用变异FP-树改进CLOSET算法  

CLOSET Algorithm Based on Variant FP-Tree

作　　者：刘迎意[1] 吴春旭[1] 沈陵峰[1] 

机构地区：[1]中国科学技术大学管理学院,安徽合肥230026

出　　处：《计算机仿真》2010年第3期98-101,共4页Computer Simulation

摘　　要：频繁闭项集提供了频繁项集的一种完整、最小表示,对频繁闭项集的挖掘是近年来数据挖掘领域研究的热点,研究人员从不同角度对算法改进以提高算法的效率。基于频繁项集中共生项集的性质,提出无须进行子集检查的频繁闭项集挖掘方法,并设计一种变异的FP-树结构,利用FP-树结构来存储结点共生项集信息,以改进CLOSET算法,算法无须遍历结果集进行闭合性检查。实验表明,在支持度阈值减小,结果集变大时,改进算法的时间增长率比原有算法小。Frequent closed item sets provide a minimal representation of frequent item sets without losing their support information. Mining the frequent closed item sets is a popular issue in data mining area in recent years. Many researchers have done a lot of work to improve the efficiency of the algorithm in different aspects. ？ In this paper, we propose a method of mining frequent closed item sets without subset checking based on the properties of co - occurrence item sets in frequent closed item sets. Moreover, we improved the CLOSET algorithm by designing a variant structure of FP - Tree to store the information of co - occurrence item sets. The improved algorithm avoids checking the closure of item sets recursively. Our experiments show that the improved algorithm outperforms CLOSET algorithm when the result set grows as the support threshold decrease.

关 键 词：数据挖掘 频繁闭项集 算法改进 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]