基于RS理论的快速属性约简求核方法  

A Rapid Approach to Compute Attribute Core of Attribute Reduction Based on RS

作　　者：代广珍[1,2] 徐超[3] 

机构地区：[1]安徽工程大学电气工程学院,安徽芜湖241000 [2]安徽工程大学电气传动与控制省级重点实验室,安徽芜湖241000 [3]安徽大学电子科学与技术学院,安徽合肥230039

出　　处：《计算机技术与发展》2011年第4期133-135,140,共4页Computer Technology and Development

摘　　要：粗糙集是用来处理不确定、不完备数据的重要工具之一。属性约简是粗糙集理论研究的一个重要内容,属性核则是属性约简所涉及的一个重要概念,对简化属性约简具有不可替代的重要性。文中指出属性约简的作用,及其涉及到的一个重要概念——属性核的概念和重要性。分析了目前常用求属性核方法,大都建立在内存中,需要构造差别矩阵,存在时空复杂度较大的不足。提出了一种无需建立差别矩阵的简单求属性核方法,并通过实例验证了正确性。Rough set theory is a new mathematical tool to research imprecise and incomplete data.Attribute reduction refers to delete unrelated or unimportant knowledge with keeping the ability of knowledge classification.Attribute core is an important concept of attribute reduction and has irreplaceable significance to simplify attribute reduction.Attempts to point out the role of attribute reduction and analyze an important conception concerned and its importance—attribute core.Methods used now almost stay in memory,and need to create discernibility matrix,so that it may lead to some demerits as time and space complex.Based on these defaults,an easier way is given to compute attribute core without creating discernibility matrix and verified by an example.

关 键 词：属性约简 属性核 差别矩阵 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]