节点文献
关联规则挖掘EARM算法及其应用的研究
A Study of the Association Rule Mining EARM Algorithm and It’s Application
【作者】 张世华;
【导师】 郭剑毅;
【作者基本信息】 昆明理工大学 , 模式识别与智能系统, 2002, 硕士
【摘要】 本课题通过对关联规则挖掘及其经典Apriori算法的深入研究,针对类Apriori算法的效率瓶颈,提出了一个高效的关联规则挖掘算法,即EARM(Efficient Association Rule Mining)算法。 一般的类Aprioir算法会产生非常庞大的候选项集,对挖掘的效率是一个沉重的负担,并且类Apriori在每一阶段循环都需要进行重复数据库存取确认,这对系统而言也是一个很大的负担。另外只用支持度作为频繁项集产生的度量指标,并未真正考虑实际交易数量和不同商品的消费一定会产生不同的获利情况等显著性的问题;针对这些不足EARM算法很好的给予了改进,使关联规则的挖掘效率得到了改善,并将EARM算法在CRM系统中给予应用。本论文将对EARM算法及其应用进行详细的研究,分析和验证。 在本论文中,首先对关联规则挖掘及其算法,历史进行了一定的介绍,从中我们可以看出,虽然关联规则挖掘产生的时间并不是很长,但它的发展和在现实生活中的应用却很快。接着详细介绍了类Apriori算法的原理和特点,在这样的基础之上,提出了高效的EARM算法。接下来,本论文对EARM算法的原理、结构和实现进行了详细的分析,并对EARM算法与其他类Apriori算法的执行效率进行比较。接着给出了EARM算法在客户关系管理系统中的应用实现,使理论研究上升到实际的应用。
【Abstract】 To solve the problem of efficiency bottleneck of Apriori-like algorithm, we propose a high-efficient association rule mining algorithm, namely EARM( Efficient Association Rule Mining) algorithm via further study of association rule mining and typical Apriori algorithm.General kinds of Apriori-like algorithm always produce huge number of candidate itemset, and it’s a grave burden to mining efficiency. At the same time Apriori-like algorithm needs repeated confirming of the database FETCH in every stage circulation. That is also a great burden for the system. In addition, only use the threshold of support for measurement of the frequent items, but not really think that the actual trade amount and the consuming of different commodity will produce the different payoff and others; EARM algorithm make good improvement for this shortage, and perfect for the mining efficiency of the association rule, and apply in the CRM system well. This paper will research , analyze and verify in detail on EARM algorithm and it’s application.In this paper, introduce the association rule , it’s algorithm and it’s history firstly, we can know that although the birth time is not long of the association rule, it’s development and application in life is speedy. Secondly, introduce the theory and character of the Apriori-like algorithm, based on this, propose the efficient EARM algorithm. Then this paper analyze the theory , structure and application of the EARM algorithm at length , and compare the implement’s efficiency of the EARM with the Apriori-like algorithm. Then give the application’s realization of the EARM in Customer Relation Management system to make the theory upgrade practical application.
【Key words】 Date mining; Association rule; Frequent itemset; Candidate itemset;
- 【网络出版投稿人】 昆明理工大学 【网络出版年期】2004年 01期
- 【分类号】TP311.131
- 【被引频次】1
- 【下载频次】75