根据自反模糊关系,将知识粒度的概念推广为模糊知识粒度。考虑传统模糊粗糙集的粗糙性度量和相似性度量,忽略了模糊集的粗糙近似处于不同知识粒度背景中这样一个重要因素,结合模糊知识粒度的计算,提出了模糊粗糙集的粗糙性度量和相似性度量的新方法。最后,在一个实际的模糊信息系统中,给出了基于模糊知识粒度的知识约简算法。
In this paper, the concept of knowledge granularity was extended to fuzzy knowledge granularity based on reflective fuzzy relation. In combination with the fuzzy knowledge granularity calculation, the new measurements of roughness and similarity of the fuzzy rough sets are proposed by taking into account their traditional measurements but ignoring such an important factor that the roughness in two different fuzzy sets may be approximated in two different levels of knowledge granularity. Finally, from the viewpoint of the fuzzy knowledge granularity, the algorithm of knowledge reduction is presented in a practical fuzzy information system.