面向商务智能的思维主题发现

项目名称： 面向商务智能的思维主题发现

项目编号： No.71272161

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 管理科学

项目作者： 高学东

作者单位： 北京科技大学

项目金额： 55万元

中文摘要： 商务智能数据挖掘应用过程中"业务理解"和"数据理解"阶段任务（即确定数据挖掘分析主题及任务）缺乏工程化解决方法,阻碍了数据挖掘的工程化应用进程,影响商务智能工程化实践。 本项目针对数据挖掘工程化应用过程中上述障碍问题，研究面向业务决策人员的思维建模及分析主题发现技术。旨在模拟决策人员思维模式，给出主动确定数据挖掘主题、任务及相关数据组织策略的结构化方法。研究内容包括： （1）思维单元的结构化建模技术，形式化表示思维单元及其之间的联系。 （2）思维分析主题发现方法，包括思维序列抽取、思维序列聚类和分析主题特征序列提取等子问题。 （3）思维分析主题逻辑结构识别方法。 （4）思维分析主题漂移检测及更新方法，检测思维分析主题漂移现象，并对其进行增量式更新。

中文关键词： 商务智能；数据挖掘；思维分析主题；数据挖掘任务；概念对

英文摘要： In the application process of data mining technology for business intelligence, 'business understanding' and 'data understanding' are accomplished by the data-mining analysts, who will try to understand the intelligent data analysis demands from business prospective, and translate them into data mining theme and tasks. The intentional manipulation causes a blockage in the process of engineering for data mining application and the normalization of business intelligence. The project focuses on the obstacle in data mining engineering application process, and research on thoughts modeling and analytical theme discovery technology. The aim of the research is to propose a systematic method for automatically discovering data mining analytical theme, task and data collection strategy, by simulating the thinking pattern of decision makers. The research includes following issues: (1) Thought unit modeling technology, which raises formally representation for thoughts unit and the relationship between them. (2) Thought analytical theme discovery method, which includes thought sequence identification algorithm, thought sequence clustering method, and extraction method of feature sequence for thought theme etc. (3) Logical structure analysis for thought analytical theme. (4) Thought analytical theme drift detection and update

英文关键词： BI；DM；thinking analysis subject；data mining tasks；concept pairs