论文部分内容阅读
经典的粗糙集理论提出知识是有粒度的,但它没有量化信息粒度所表示的信息量。本文定义了知识粒数,知识粒的微粒数,平均微粒数,近1系数,知识粒度的概念;提出了知识粒度的量化计算方法;提出了一种基于知识粒度的属性约简算法以避免选择约简子集的盲目性;给出了时间复杂度证明;通过海上交通事故的决策实例证明所提出的粒度计算方法是可行有效的。
Classical rough set theory suggests that knowledge is granular, but it does not quantify the amount of information represented by information granularity. This paper defines the concept of knowledge granularity, the number of particles, the average number of particles, the coefficient of near 1, and the granularity of knowledge. The paper proposes a quantitative computing method of knowledge granularity and proposes a attribute reduction algorithm based on knowledge granularity to avoid selection The blindness of reduction subset is given. Proof of time complexity is given. It is feasible to validate the proposed granularity calculation method through examples of decision-making of maritime traffic accidents.