《关联规则在服装销售行业的应用研究》-毕业论文设计（学术).doc

工程硕士学位论文 关联规则在服装销售行业的应用研究 AAA 北京工业大学 摘 要 数据挖掘（Data Mining）是近十几年中迅速发展起来的交叉学科，它融合了数据库、人工智能、统计学等多个领域的理论和技术，搭建了上述理论研究与实际应用间连接的桥梁,应用范围广泛。本文提出了利用数据挖掘技术对某服装企业的销售数据进行挖掘分析研究，以求找到大量数据背后潜在的信息；并给出了研究工作的思路和方法；得出了部分对企业有意义的信息并积累了一定的工作经验。 本文首先概述了数据挖掘理论和发展，以及主要的数据挖掘技术；然后，基于某服装企业02年到07年的销售数据，重点研究了关联规则在销售数据上的应用问题。 主要内容包括: 首先对原有数据进行了预处理 ；其次运用了Apriori算法产生了大量的关联规则；再次使用了闭频繁项集和最大频繁项集概念结合Apriori算法来解决规则冗余问题；最后应用作用度（Lift）对所挖掘出的规则进行评估，去掉负关联规则后，用服装行业知识对剩余的正关联规则进行评判筛选，找出对服装企业有用的规则。 试验结果表明由于使用的数据集自身的问题，应用闭频繁项集消除规则冗余的效果并不明显，而使用最大频繁项集理论在消除冗余规则方面效果十分显著。并且通过应用作用度（Lift）对规则的评估后得出部分对服装企业有意义的规则，对该行业今后的生产和销售有一定的指导作用  Abstract Data Mining is the most popular intercross subject in the last decade. It integrates several areas of theories and technologies, such as the Database, Artificial Intelligence and Statistics, becomes the bridge to connect the theory research and practical applications, and is wildly used. This thesis illuminates utilizing the data mining technology for one costume enterprise to process the data mining research on its sales data, in order to find potential information behind the enormous data; it also presents the idea and method for how to process the research work. From the research described in the thesis, some significative information for the enterprise are achieved, and we also get certain research experience on the area. This thesis first summarizes data mining theory, its evolution and its primary data mining technology. Then it focuses on the research of applying the data mining technology of association rules to one costume enterprise’s sales data, which are collected from financial year 2002 to 2007. The following main contents of the research are included in the thesis: first the original data is pre-processed, then the classic Apriori algorithm is applied on the data to generate plentiful association rules. In order to eliminate a mass of redundant rules, the theory of frequent closed itemsets and the maximal frequent itemsets are utilized with the Apriori algorithm