基于二维数组和十字链表的Apriori算法  

Apriori Algorithm Based on Two Dimensional Arrays And Orthogonal List

作　　者：陈衡[1] 刘玉文[2] 

机构地区：[1]淮北职业技术学院,安徽淮北235000 [2]蚌埠医学院卫生管理系,安徽蚌埠233000

出　　处：《德州学院学报》2017年第2期63-67,共5页Journal of Dezhou University

基　　金：安徽省高校质量工程项目(2015zjjh051);淮北职业技术学院自然科学重点项目(2016-A-4)

摘　　要：关联规则挖掘的核心是寻找频繁项集,其有两个技术瓶颈:(1)容易生成大量无效候选项集;(2)需要多次扫描数据库.基于二维数组和十字链表的Apriori改进算法,只需扫描数据库一次,对频繁(k-1)-项集进行分组,在分组的基础上生成候选k-项集.另外,事务数据库以十字链表形式表示,提高了候选项集计数效率、降低了内存使用空间.实验结果表明,Apriori改进算法在运行效率上比其他算法有一定程度的提高.Finding frequent itemsets is the core of association rule mining, It has two technical bottle- necks, One is to scan the database many times, Another is to generate a large number of invalid candidate itemsets. The paper presents an improved Apriori algorithm based on Two--dimensional arrays and Or- thogonal List, the algorithm scans the database only once, and generate k--candidate itemsets based on the frequent （k- 1） -- itemsets packet. In addition, the transaction database express by cross list, the candidate itemset counting efficiency is improved, and memory usage space is reduced. The experimental results show that the efficiency of the improved algorithm is higher then orthers.

关 键 词：APRIORI算法 候选项集 频繁项集 二维数组 十字链表 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]