基于模糊邻域粗糙集的信息系统不确定性度量方法

Uncertainty measurement method for information system based on fuzzy neighborhood rough set

邻域粗糙集和模糊粗糙集是粗糙集理论中处理数值型数据的两种重要模型.在数值型信息系统中融合两者在不确定性度量方面的优越性,首先引入了模糊邻域粗糙集模型,并在该模型上定义了模糊邻域粗糙度的概念.模糊邻域粗糙度是通过粗糙集的边界域来度量信息系统的不确定性,为了达到更为全面的度量效果,在模糊邻域粗糙集模型中定义了模糊邻域粒结构,并基于该粒结构提出了模糊邻域粒度的概念,模糊邻域粒度是对信息系统分类能力的一种度量.最后,通过将两种度量方法进行结合,提出了一种基于模糊邻域粗糙集的混合不确定性度量方法,并从理论上证明其有效性.实验结果表明,所提出的混合度量方法综合了两种单独度量方法的优点,在数值型信息系统中具有更好的度量效果,因此所提出的不确定性度量方法更具有一定的优越性.

The neighborhood rough set and the fuzzy rough set are the two kind of important models for processing numeric data in rough set theory.In the numerical information system,combining with the superiority of the neighborhood rough set and the fuzzy rough set in terms of uncertainty measurement.The model of fuzzy neighborhood rough set is firstly introduced in this paper,and the conception of fuzzy neighborhood roughness is defined on the model of fuzzy neighborhood rough set.The fuzzy neighborhood roughness measures the uncertainty of information system through the boundary region of rough set,which is aimed to obtain more comprehensive measurement effect.And then,the fuzzy neighborhood granular structure is defined on the model of the fuzzy neighborhood rough set,and the concept of fuzzy neighborhood granularity is proposed based on the fuzzy neighborhood granular structure,and the fuzzy neighborhood granularity is a measure of the classification capacity for information system.At last,the method of hybrid uncertainty measurement based on fuzzy neighborhood rough set is proposed through combining two measurement methods,and which is theoretically proved effective.Experimental results show that the proposed method of hybrid measurement integrates the advantages of the two separate measurement methods,which has better effect of measurement.Therefore,the proposed method of uncertainty measurement has more certain superiority in this paper.