不完备集中有序信息系统的属性约简  

Attribute reduction of incomplete ordered concentration information systems

作　　者：金玲玲[1] 汪文俊[1] 

机构地区：[1]海南师范大学数学与统计学院,海南海口571158

出　　处：《辽宁工程技术大学学报（自然科学版）》2016年第11期1316-1321,共6页Journal of Liaoning Technical University (Natural Science)

基　　金：海南省自然科学基金项目(20156233;114005)

摘　　要：针对现有集中有序关系模型对不完备信息处理的局限性,利用先验概率对属性取值进行估计,重新定义了对象间的优势度,对集中有序关系进行扩展.引入知识粒度的概念,以知识粒度的变化作为属性重要性的度量,提出了适合于不完备集中有序信息系统的属性约简算法.研究结果表明:建立了不完备信息系统中趋于某个标准值的一种偏好关系,约简算法切实可行.Attribute values of information system are not all ordered according increasing preferences or decreasing preferences.Preference order based on a standard value is called ordered concentration.In view of limitation of incomplete information processing in existing ordered concentration relation model,prior probability method in estimating attribute values was used,dominance between objects were redefined and the ordered concentration relation was extended.Including the concept of knowledge granularity,which was used to measure the importance of attributes,the algorithm of attributes reduction was proposed to make it suitable for the incomplete ordered concentration information system.The research demonstrates that a kind of preference relation of tending to a standard value was established in incomplete information system,and the reduction algorithm is feasible.

关 键 词：不完备 集中有序 标准值 概率 属性约简 知识粒度 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]