基于相似度的粗糙集近似算子快速求解  

Fast Solution of Similarity-based Rough Set Approximate Operators

作　　者：仲兆满[1] 刘宗田[1] 周文[1] 付剑锋[1] 

机构地区：[1]上海大学计算机工程与科学学院,上海200072

出　　处：《小型微型计算机系统》2010年第1期151-154,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60575035)资助;上海市重点学科建设项目(J50103)资助;上海高校选拔培养优秀青年教师科研专项基金项目(shu-07027)资助

摘　　要：由于经典粗糙集只能处理精确分类问题,基于相似度的粗糙集模型被提出并用于解决不完备信息系统的相关问题.粗糙集通过近似算子对某一给定的概念进行近似表示,科学的求解这些算子对粗糙集理论的发展具有重要意义.本文提出一种新的近似算子快速求解方法,分析证明了所提快速方法比经典方法具有更高的求解效率.文章定义了元素覆盖度、集合覆盖度等概念,使用覆盖度等价关系可以将覆盖粗糙集转化为经典粗糙集,从而简化覆盖粗糙集的相关问题的解决.Similarity-based rough set models are put forward to solve incomplete information systems because classical rough set is only used to deal with precise classification. Rough set is used to approximately represent a certain concept by approximate operators, and getting these operators effectively is of great significance for the development of rough set theory. A new fast solution of getting approximate operators is presented, and compared with classical method, the method proposed has higher efficiency. The covering degree of elements and sets are defined, and applying equivalent relation of covering degree can translate covering rough sets into classical rough set easily, so that simplifying the theory of covering rough sets.

关 键 词：经典粗糙集 覆盖粗糙集 近似算子 快速求解方法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]