面向对象的属性约简系统框架的设计与实现  

DESIGN AND REALIZATION OF ATTRIBUTE REDUCTION SYSTEM FRAMEWORK BASED ON OBJECT-ORIENTED TECHNOLOGY

作　　者：侯海平[1] 王锋[1] 宋旭东[1] 孙辉[1] 

机构地区：[1]大连交通大学软件学院,辽宁大连116028

出　　处：《计算机应用与软件》2008年第12期27-29,共3页Computer Applications and Software

基　　金：国家科技部十五科技基础工作专项资金(2002DEA20012)

摘　　要：传统的Rough集理论在数据挖掘中的应用大多是通过结构化的方法来实现的,但是这种方法存在着以下不足:数据模型结构性不强、软件复用率低、可扩展性差等。为了解决这些问题,将面向对象技术引入到Rough集理论的研究中。首先,提出了一种面向对象的属性约简系统框架;其次,设计了一种属性约简的对象模型;最后,基于该框架对UCI数据进行了验证。实验结果表明,该框架能有效地对数据集进行约简。此外,该框架结构简单、具有良好的扩展性。提出的对象模型同样适用于其它的数据挖掘系统。Though traditional Rough set theory applications in data mining are mostly based on structured method, it can possibly involve issues such as poor structural integrity of data model, low software reuse and lack of extensibility, To solve these problems, in this paper we apply the object-oriented technology in the research of Rough set theory. First, we propose a system framework of object-oriented attribute re- duction. Next, we design an object model of attribute reduction. Finally, we carry out a verification experiment on UCI data as source data based on the framework. The results show that the dataset is reduced effectively with this framework. In addition, this framework is simple and has good extensibility. The object model provided in this paper is also suitable for other data mining system.

关 键 词：ROUGH集 属性约筒 面向对象 模型 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论] TP311[自动化与计算机技术—计算机科学与技术]