【作者】 曹丹阳；

【导师】 李晋宏；

【作者基本信息】 北方工业大学 ， 计算机应用技术， 2006， 硕士

【摘要】 数据仓库和数据挖掘是数据库研究、开发和应用最活跃的分支之一,也是决策支持系统的关键因素,数据仓库是一个支持管理决策过程的、面向主题的、随时间而变的数据集合,它是集成的,也是稳定的。数据挖掘是采用人工智能的方法对数据库和数据仓库中的数据进行分析、获取知识的过程。它们的结合能更好地为企业或有关部门不同范围的决策分析提供有力的依据。 纵观以往的教学管理系统,多半是OLTP系统,缺乏综合分析、辅助决策的能力;并且对其历史积累的海量信息中隐含知识的利用无能为力。对教学管理进行分析是教学评估的重要手段,采用数据挖掘技术对教务数据进行多层次、多角度的分析与挖掘,利用挖掘结果辅助教学决策是保证教学质量、提高学生素质的必然要求。 本文主要探讨了基于数据仓库的数据挖掘技术的基本理论和实施方法,探索了数据挖掘分类方法和聚类方法。结合教务系统,改进分类方法中的决策树算法并应用于英语四级成绩分析,实现聚类方法中的K-平均算法和k-中心点算法并应用于学生毕业情况分析,从而实现了基于教务的数据挖掘系统。 本文首先从决策分析需求出发构建教务系统数据仓库;接着对数据进行预处理并改进概念分层方法,使概念分层更适合教务系统数据挖掘;而后对预处理后的数据以图、表的形式分析统计;最后利用改进的数据挖掘算法对数据进行挖掘,得出有效结果并运用于实践中。 通过在教务系统中的具体挖掘实践,得到了许多有价值的信息,这些知识在帮助学校更好地进行学生的培养,对学生表现情况的掌握以及课程的安排等方面无疑具有重要的指导意义。更多还原

【Abstract】 Data warehouse and data mining are two of the most active branches of database studying, developing and application, and also the key factors of DSS. Data warehouse is a decision supporting, subject-oriented, integrated, stable and time-dependent collection of data;Data mining is to analyze data and acquire knowledge from database and data warehouse using the method of artificial intelligence. The binding of them will provide a strong basis of decision analyzing for enterprises and related departments.Most of the previous teaching management systems are the On-Line Transaction Processing (OLTP) systems that have no ability of synthetic analysis, decision support, and the utilization of hidden knowledge from vast history information. Analysis of teaching management is an important way to teaching evaluation. It is necessary to guarantee the quality of teaching and improve the stuff of students analyzing the data made in processes of tests and teaching and then supporting the teaching with the results of analysis.In this paper, I introduce the fundamental theory and technologies of data mining based on data warehouse, and research method of classify and cluster in data mining. Combined with the practice in educational administration system, I has improved decision tree algorithm in the classify algorithm, and applied it to CET-4 grade analysis, and besides, I has realized the k-means and k-mediods algorithm in the cluster algorithm, and applied it to circs analysis of a graduate of a college. I put emphasis on the design and implementation of data mining system for educational administration data.In the system, I mainly realize the following functions: to construct suitable data warehouse based on decision and analysis requirement about educational administration system, to pretreat data and improve method of conception level, to analyze pretreated data at chart and table, to mine data by improving algorithm, to conclude on effectual results and put its into practice.By applying the method to educational administration system we obtain much valuable knowledge which is helpful for college education.更多还原