两种改进的关键规则挖掘算法研究与应用

发布时间：2018-01-12 19:08

本文关键词：两种改进的关键规则挖掘算法研究与应用　出处：《合肥工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着数据库技术的快速发展和应用,累积了大量的数据,如何有效、合理的运用这些数据,发掘数据背后隐藏的知识成为了人们关注的重点。传统的统计数据分析已难堪大任,而数据挖掘作为分析数据的一种新的工具受到越来越多的重视,成为信息技术的热门研究领域。关联规则作为数据挖掘的一个重要的分支,也得到了长足的发展,在很多方面已经取得了重要的成果,但在面对海量数据情况下仍然面临着很多挑战。随着关联规则的发展,学者提供了各种关联规则挖掘算法,比较经典的包括Apriori算法、Fp-Growth算法、DHP算法等。但是结合计算机技术和实际应用场景很多算法都有较大的改进空间,因此本文在分析了国内外关联规则研究现状的基础上,提出了两种改进的关联规则挖掘算法。本文的主要工作如下。(1)对数据挖掘和关联规则的基本理论进行了分类总结。并详细阐述了Apriori和DHP两种经典的关联规则挖掘算法。(2)提出了改进的Apriori算法——DecBit Apriori算法,DecBitApriori算法将事物数据库转换成十进制数数据库,然后使用与位运算计算候选集的支持度。最后通过实验验证了DecBitApriori算法在运行效率上的提高。(3)提出了改进的DHP算法——RBTDHP算法,RBTDHP算法使用红黑树数据结构处理DHP算法散列过程中的冲突,可以让所有的候选集单独计数,避免了DHP算法需要重复扫描数据库得到候选集的支持度。(4)基于超市实际购物数据,使用DecBitApriori算法挖掘关联规则,然后根据挖掘出的关联规则结果提出了超市商品摆放的相关建议。
[Abstract]:With the rapid development and application of database technology, accumulated a large amount of data, how to effectively use these data reasonably, explore the knowledge behind the data become the focus of attention. The traditional statistical data analysis and data mining has great embarrassment, as a new tool of data analysis has received more and more attention and become a hot research field of information technology. As an important branch of data mining association rules as, also obtained the considerable development, the important results have been achieved in many ways, but in the face of massive data situation still faces many challenges. With the development of association rules, scholars provide a variety of mining association rules the algorithm, including Apriori algorithm, Fp-Growth algorithm is the classic DHP algorithm. But the combination of computer technology and practical application of the scene many algorithms have changed greatly In space, therefore, based on the analysis of the domestic and foreign research status of association rules, this paper proposes two improved association rule mining algorithm. The main work of this paper is as follows. (1) are classified and summarized the basic theory of data mining and association rules. And expounds the association rules Apriori and DHP two classic the data mining algorithm. (2) proposed the improved Apriori algorithm, DecBit Apriori algorithm, DecBitApriori algorithm will convert the decimal number database transaction database, and then use the support and computing candidate sets. Finally we validate the DecBitApriori algorithm to enhance the efficiency in the operation. (3) proposes an improved DHP algorithm. RBTDHP algorithm, RBTDHP algorithm uses conflict red black tree data structure DHP algorithm hash process, can make all the candidate set separate count, to avoid the need to re DHP algorithm The support of the candidate set is obtained from the complex scan database. (4) based on the actual shopping data of supermarkets, we use DecBitApriori algorithm to mine association rules, and then put forward relevant suggestions for supermarket's commodity placement according to the results of mining association rules.

【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.13

【参考文献】