导　　师： 许龙飞

学科专业： H1202

授予学位： 硕士

作　　者： ;

机构地区： 暨南大学

摘　　要： 本文介绍了一个基于j2ee框架的多层数据挖掘(datamining)原型系统。它利用rational公司的rose工具进行系统的uml面向对象可视化建模，并使用sun公司的javawebstart进行客户端应用程序的开发。提高了系统的可靠性、可重用性和可维护性。在数据挖掘原型系统中，实现了关联分析的apriori算法、分类的id3决策树算法、c4.5的悲观估计决策树算法和c4.5决策树的消除误差修剪算法(reduced-errorpruning)。为开发其它数据挖掘应用系统提供了一个有价值的实例。 In this thesis, a multi-layer data mining prototype system based on the popular J2EE framework is introduced. The prototype is visually modeled using UML object-oriented technology of Rational Company. The client application of data mining is developed conveniently using Java Web Start technology. The reliability, reusability and maintainability of the system are enhanced because of above technologies. In the data mining prototype system, Apriori algorithm of association rules mining, ID3 algorithm of decision tree classification, C4. 5 pessimism estimate algorithm of decision tree classification and C4. 5 reduced-error pruning algorithm of decision tree classification are realized. This thesis provides a valued sample for developing other data mining application systems.

关 键 词： 数据挖掘 面向对象技术 系统设计

分 类 号： [TP311]

领　　域： [自动化与计算机技术] [自动化与计算机技术]