不完备邻域多粒度决策理论粗糙集与三支决策  

INCOMPLETE NEIGHBORHOOD MULTI-GRANULATION DECISION-THEORETIC ROUGH SET AND THREE-WAY DECISION

作　　者：刘丹 徐立新 李敬伟 Liu Dan;Xu Lixin;Li Jingwei(College of Computer Science and Technology,Henan Institute of Technology,Xinxiang 453003,Henan,China)

机构地区：[1]河南工学院计算机科学与技术学院,河南新乡453003

出　　处：《计算机应用与软件》2019年第5期145-157,共13页Computer Applications and Software

基　　金：河南省高等学校重点科研项目(19B520005)

摘　　要：多粒度决策理论粗糙集是多粒度视角下三支决策中一种重要的模型。在数值型不完备数据下建立邻域容差关系;在其基础上提出乐观和悲观的邻域多粒度决策理论粗糙集模型。为了弥补这两种模型的局限,提出平均邻域多粒度决策理论粗糙集模型,并分析相关性质以及相互关系。同时为了使所提出的邻域多粒度决策理论粗糙集适用于不完备数据环境,运用区间值的形式表示代价函数,并通过选取不同参数的方式提出一种可变三支决策。实例分析表明,该模型与方法具有一定的合理性与灵活性。Multi-granulation decision-theoretic rough set is an important model in three-way decision under multigranulation perspective.The neighborhood tolerance relation was established under the numerical incomplete data,and we proposed optimistic and pessimistic neighborhood multi-granulation decision-theoretic rough set model.To make up for these two models'limitation,we further proposed the mean neighborhood multi-granulation decision-theoretic rough set model,and analyzed the related properties and relations.In order to make the proposed model suitable for incomplete data environment,we used interval form to express cost functions,and proposed a variable three-way decision by selecting different parameters.The example analysis shows that the model and method have certain rationality and flexibility.

关 键 词：决策理论粗糙集 多粒度 邻域 不完备信息系统 三支决策 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]