针对Jensen下近似定义的局限性,提出一种新的等价类形式的近似算子表示,并将其推广到直觉模糊环境.在此基础上,将相对正域、相对约筒、相对核等粗糙集的知识约简概念推广到直觉模糊环境,提出一种直觉模糊信息系统的启发式属性约简算法.实例计算表明,该方法比Jensen的属性约筒方法更为合理有效.
To the issue of lower approximation proposed by Jensen, an improved definition of lower approximation described with fuzzy equivalent classes is proposed and extended to intuitionistic fuzzy approximation space firstly. The notion of positive region and relative reduct in classical rough set are generalized to the intuitionistic fuzzy information system and a heuristic algorithm for attribute reduction based on intuitionistic fuzzy rough set is proposed. Finally, two particular experiments verify the rationality and effectiveness of the developed algorithm.