摘要
在不完备多粒度决策信息系统中,已有多种拓展优势关系来处理多粒度粗糙集模型,然而未考虑优势关系的有序特性。所以希望在多粒度粗糙集模型中引入置信优势关系对条件属性作划分,考虑置信优势类与目标系统之间一致的定量信息。文中所研究的多粒度是基于属性集序列来对论域进行划分的多粒度。考虑不同粒度重要性的不同,讨论了置信优势关系下加权多粒度粗糙集模型的乐观、悲观多粒度粗糙集近似模型,提出了一种基于分类质量的粒度约简算法,最后通过算例验证与评估所提理论与方法的正确性与可行性。
In incomplete and multi-granular decision information systems,there have been a variety of extended dominant relations to deal with multi-granular rough set models.However,the orderly nature of the dominance relations is not considered.Therefore,this paper hopes to introduce the confidential dominance relations in the multi-granular rough set model to divide the condition attributes,and analyze the consistent quantitative information between the confidential dominance class and the target system.The multi-granular studied in this paper is based on the attribute set sequences to divide the domain.Considering the different importance of different granules,the optimistic and pessimistic approximation model of weighted multigranulars rough set based on confidential dominance relation is further discussed.A subtle reduction algorithm based on classification quality is proposed.Finally,an example is given to validate and evaluate the correctness and feasibility of the theories and methods proposed in this paper.
作者
林玉梅
方连花
郭新华
LIN Yu-mei(School of Software,Quanzhou University of Information Engineering,Quanzhou Fujian 362000,China)
出处
《长春工程学院学报（自然科学版）》
2020年第2期119-124,共6页
Journal of Changchun Institute of Technology：Natural Sciences Edition
基金
福建省教育厅中青年教师教育科研项目(JAT190920)
关键词
置信优势关系
多粒度粗糙集
粒度约简
confidential dominance relation
multi-granular rough set
granulationreduction
作者简介
林玉梅(1982-),女(汉),福建人,讲师主要研究粗糙集、网络安全。
