诱导覆盖关系下的加权多粒度粗糙集及应用

Weighted multi-granulation rough set under induced covering relation and its application

为从更多角度进一步分析多属性覆盖决策信息系统,提出了诱导覆盖关系下的加权多粒度粗糙集。分析了采用专家经验确定粒度权重的局限性,引入基于分类质量的粒度权重赋值方法,使决策结果更具客观性;将诱导覆盖关系引入加权多粒度粗糙集,给出完整的粗糙上下近似;讨论了几种度量参数之间的关系,并对相关性质和定理进行证明;通过实例分析验证了模型的有效性和实用性。

As a further exploration of the multi-attribute covering decision information system from more diverse angles,this paper proposed a weighted multi-granularity rough set based on induced covering relation.Firstly,it analyzed the limitation of granular weight’s determination depend on expert experience.In order to obtain relatively objective decision results,it adopted the granular weight assignment method based on classification accuracy.Then,it introduced the induced covering relation into the weighted multi-granulation rough set,and acquired the complete upper and lower approximations of the rough set.Further,it discussed the relationship between metric parameters and proved the related properties and theorems.Finally,the experimental results verify the validity and practicability of the model.