犹豫值决策系统的序结构及不确定性度量

Order Structure and Uncertainty Measures of Hesitation Decision System

犹豫模糊决策系统作为一种集值类型的信息系统,由于其属性值具有犹豫模糊性,能更为客观地反应涵盖在数据中的不确定性信息。在犹豫值决策系统的研究中,一个重要的研究课题是从犹豫值反映的属性下的多个信息中挖掘出不同样本之间的特殊关系。由于当前犹豫模糊决策序结构的研究存在短板,本文构造了一种优势度,用以描述犹豫模糊决策系统中样本的偏序关系。结合粗糙集理论,本文利用前述优势度进一步构造不确定性测度,用于度量属性子集保持决策的偏序能力。基于这种不确定性测度,提出一种前向贪婪的特征选择方法。最后,结合实例,给出犹豫模糊决策系统的特征选择过程。本文的研究思路和研究结果对于度量偏序关系、构造不确定性测度及修正特征选择方法有参考借鉴意义。

As a set-valued information system,hesitant fuzzy decision-making system can more objectively reflect the uncertain information contained in the data because its attribute value is hesitant fuzzy.It is a critical issue to mine the special relationship between different samples in the study of hesitant fuzzy decision system for the multiple information reflected from the hesitant fuzzy values under attributes.In view of the shortcomings of the existing technologies in the study of the order structure of hesitant fuzzy decision system,we creatively construct a dominance to measure the partial order relationship of samples in hesitant fuzzy decision system.Combined with the rough set theory,we construct an uncertainty measure to measure the ability of attribute subset to keep decision partial order further.Based on this uncertainty measure,a forward greedy feature selection method is proposed.Finally,the feature selection process of the hesitant fuzzy decision system is given with an example.This paper describes an innovative partial order,builds an uncertainty measure,and repairs the feature selection method.The analytical methods and suggestions given in it are beneficial to other researchers.