数据挖掘关联规则Apriori算法的一种新改进
首发时间:2009-10-12
摘要:关联规则算法的研究在数据挖掘算法的研究中占有相当重要的地位。关联规则算法的核心是Apriori 算法,但随着对关联规则研究的深入,它的缺点也暴露出来了。Apriori算法有两个致命的性能瓶颈:多次扫描事务数据库,需要很大的I/O负载;产生庞大的候选集。所以Apriori 算法仍有需要改进的地方。本文提出了一种新的基于矩阵的改进Apriori算法, 充分利用了内存空间,大大减少扫描数据库的次数,多次试验的结果表明该方法有效提高了大型数据库的使用效率。
关键词: 改进Apriori算法 关联规则 频繁项目集
For information in English, please click here
Further Improvement in Apriori Algorithm of Association Rules in DM
Abstract:Association rule algorithm is quite important when we research on the data mining. The core algorithm of association rule is Apriori. But with the further association rule research going on smoothly, the weakness of Apriori exposes to us. The Apriori has two fatal performance bottlenecks: lots of I/O load is needed because of frequently scanning transaction database; lots of candidate itemset are produced. Therefore the Apriori algorithm needs improving. This article proposes one new algorithm, which makes full use of space, and the number of scanning database drop off sharply. Many experiments indicate that the method mentioned above greatly increase the service efficiency of large-scale database.
Keywords: Improved Apriori arithmetic Association Rules Frequent Items
基金:
论文图表:
引用
No.3573842880312553****
同行评议
共计0人参与
勘误表
数据挖掘关联规则Apriori算法的一种新改进
评论
全部评论0/1000