在经典粗糙集中通常运用不可区分关系对知识进行粒化.为解决知识粒化过细带来的分类和决策困难,文章在程度不可区分关系的基础上,讨论了各种可区分关系和不可区分关系的演化机制,提出了基于序信息系统的改进程度可区分关系,进一步刻画对象属性值间的差异程度,最后建立了一种基于程度可区分关系的概率粗糙集模型.%In general,the knowledge granularity can be induced by indiscernibility relation based on typical rough sets. In order to overcome the difficulty of classification and decisions caused by more refined knowledge granularity, this paper discussed the evolution mechanism under various indiscernibility and discernibility relation,and proposed the improved grade discernibility relation based on ordered information systems to characterize the difference of grade be-tween object's attribute values based on grade indiscernibility relation. Finally,it introduced a rough probabilistic sets model based on grade discernibility relation.
展开▼