粗糙集理论中一种属性离散化算法  

An Algorithm for Discretization of Condition Attributesin Rough Set Theory

作　　者：安利平[1] 仝凌云[2] 

机构地区：[1]南开大学国际商学院,天津300071 [2]河北工业大学管理学院,天津300130

出　　处：《河北工业大学学报》2002年第3期39-43,共5页Journal of Hebei University of Technology

摘　　要：针对定量属性离散化制约粗糙集理论应用这一关键问题,利用分级聚类法和粗糙集理论中依赖度的概念,提出了一种对决策系统中条件属性进行离散化的增类减类算法,实现了决策系统的属性约简.该算法易于理解,计算简单,以实例说明了算法的合理性和有效性.The discretization of quantitative attributes is the critical factor that limits the application of rough set theory. In this paper, using grade classification method and the concept of dependency degree, an algorithm is presented for discretization of condition attributes in a decision system. Furthermore, attribute reduction can be achieved by using the algorithm. The algorithm is constructed with good understandability and simple computation. In the end, an example illustrates that the algorithm is reasonable and effective.

关 键 词：粗糙集 知识表示系统 属性离散化 属性选择 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]