粗糙集的核心概念在映射观点下的研究  

Study on Core Concepts of Rough Sets in Mapping View

作　　者：蒋开平[1] 胡彩萍[1] 段志锋[1] 刘超[1] 

机构地区：[1]江西师范大学计算机信息工程学院,江西南昌330022

出　　处：《计算机技术与发展》2009年第7期133-136,共4页Computer Technology and Development

基　　金：国家973前期研究专项(2007CB316505)

摘　　要：经典粗糙集的一些核心概念的定义基础是划分。目前,粗糙集的属性约简算法大部分都是基于以这种方式定义的相关核心概念。理论分析和实验表明,这类属性约简算法具有比较明显的低效性。在深入分析这些核心概念的基础上引入了映射的概念,给出了基于映射观点的有关核心概念的另一种形式的定义,并从理论上论证了这种定义方式的相关结论。论证结果表明这种定义方式在属性约简算法中有效。The definition of some eore concepts in classical rough sets based on partition. Now, most of attribute reduction algorithms are based on this type of core concepts. Theoretical analysis and experimental results show that this type of reduction algorithm has evident inefficiency. Introduce the mapping concept under making a deep study of these core concepts. Proposing and proving another type of definition and some conclusions about related core ooncepts based on the mapping concept. Theoretical analysis shows that this type of definition about core concepts is efficient in attribute reduction algorithm.

关 键 词：映射 属性核 正区域 属性依赖 属性重要度 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]