摘要
粒计算是一种基于问题概念空间划分的新的智能计算理论和方法,目前在国际上逐步得到了人工智能有关研究人员的重视。模糊粒度模型、粗糙集粒度模型、邻域系统下的粒计算模型、商空间模型、相容粒度空间模型是目前几种常用的粒计算模型。基于粗糙集理论的粒度模型,通过决策信息系统的粒子空间中各粒子的推理,给出了决策信息系统中核属性计算方法;在此基础上,提出了决策信息系统属性约简的计算方法;通过实例验证了该方法的有效性。
Granular Computing （GrC） is a new theory and method of intelligence computing based on the problem concept space partition. In recent years, GrC has attracted a considerable attention. The fuzzy granularity model, rough granularity model, neighborhood granular model, entropy space granular model and the tolerance relation granular model are common GrC models. In this paper, an approach of computing core attributes based on granular computing model in rough set theory form decision information systems was proposed. On the basis of the core attributes, a method of attribute reduction form decision information systems was developed. An example illustrates the efficiency of these approaches.
出处
《重庆邮电大学学报（自然科学版）》
北大核心
2010年第5期652-655,共4页
Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)
基金
四川省教育厅科研基金(09ZC079)
四川师范大学重点科研基金~~
关键词
粗糙集
粒计算
核属性
属性约简
rough set
granular computing
core attributes
attributes reduction
