摘要
在不完备信息系统中,针对用区间值表示一个未知参量时,整个区间内取值机会被认为是均等的,得到的结果可能会产生过大误差的问题,将三角模糊数引入到决策粗糙集中,提出了一种基于不完备信息系统的三角模糊数决策粗糙集。首先,定义了一种描述不完备信息的相似关系;然后,针对不完备信息系统中的缺失值,利用三角模糊数来获取损失函数,构建了三角模糊数决策粗糙集模型;实例表明,本文提出的方法不仅能够弥补用区间数表示的不足,而且可以突出可能性最大的主值,从而减少分类误差。
Aiming at the problems that when using an interval value to represent an unknown parameter in an incompleteinformation system, the opportunity to obtain the value over the whole interval is considered to be equal, butthe result may cause an over-large error. In order to solve this problem, a triangular fuzzy number was introducedinto decision-theoretic rough sets, and a triangular fuzzy decision-theoretic rough set under incomplete informationsystems is proposed. Firstly, a new similarity relation was defined to describe incomplete information systems.Then, in view of the missing values, a model of triangular fuzzy number decision-theoretic rough sets was constructedto obtain the loss function. Finally, examples show that the proposed method not only makes up for deficiency inrepresentation of the interval value, but also highlights the main value most likely to reduce the classification error.
出处
《智能系统学报》
CSCD
北大核心
2016年第4期449-458,共10页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金项目(61370169,61402153)
河南省科技攻关重点项目(142102210056,162102210261)
河南省高等学校重点科研项目(16A520057)
关键词
不完备信息系统
区间值
三角模糊数
决策粗糙集
incomplete information system
interval value
triangular fuzzy number
decision-theoretic rough sets