基于粗糙集的多维数据挖掘的研究与应用-计算机软件与理论专业论文.docxVIP

基于粗糙集的多维数据挖掘的研究与应用 基于粗糙集的多维数据挖掘的研究与应用 巾山大学硕士学位论文 Title：The research and application ofmulti-dimensional data mining based on rough set Major：Compmer Software and Theory Name：Shuping Li Supervisor：．Jirong Gao Abstract As the development ofdata warehouse technology and OLAP technology，People proposed Multi—dimensional data mining technology which was integrated by OLAP technology and the data mining technology．Multi—dimensional data mining earl enable the users to choose the data丘om any part of the data set，So the users can analyze the data in the different levels．This paper unions rough set and decision tree technology to research multi-dimensional data mining． The rough set and decision tree technology both have Merit and Shortcoming，SO we study with the two technologies．When the data set only has discrete value attribute，we can use rough set technology to process the attribute reduction．This paper proposes an attribute reduction algorithm based on the information entropy．This algorithm enhances the efficiency ofattribute reduction．Based On the attribute reduction algorithm，the data mining algorithm which unions rough set and decision：tree technology is proposed．This algorithm reduces the scale ofthe decision tree and enhance the efficiency ofdata mining． Finally using the data mining algorahm which unions rough set and decision tree technology，this paper has designed and developed a multi-dimensional data mining system．We develop the system based on sqbserver analysis seEver． Keyword：multi—dimensional data mining，rough set，decision tree，information entropy II 中山大学硕士学位论文 中山大学硕士学位论文 基于粗糙集的多维数据挖掘的研究与应用 引 言 随着现代信息技术的迅猛发展，全球掀起了信息化浪潮。信息产生的渠道越 来越多，信息更新的频率日益加快，各行各业均产生了数以亿计的数据库。用户 需要从大量的业务数据中探索业务活动的规律性、市场的运作趋势，并从中得到 市场竞争所需的决策信息。但利用关系数据库中的数据进行决策分析时往往会产 生许多麻烦，所查询的数据信息不是缺乏分析能力，就是因各信息系统之间相互 隔离、结构各异，使信息分析及决策人员不能据此对某一主题的相关数据进行多 角度的比较、分析，得出科学的分析结果。随着需求的变化发展，一种新的信息 处理技术——数据仓库技术应运而生。数据仓库∽】将来自各个数据库的信息进 行集成，从事务