粗集理论中不相容规则的一种修正算法

A correcting algorithm of incompatible rules in rough set theory

粗集理论是继概率论、模糊集、证据理论之后的又一种处理模糊和不确定性知识的数学工具。针对粗集理论在实际应用中常常出现的条件属性相同而决策属性不一致问题,提出了一种修正不相容规则的算法。即从相对错误分类率概念出发,在力求保持原知识分类能力的前提下,通过比较不相容规则的相对错误分类率大小,从而对不相容规则进行修正的算法。并通过雷达分类实例验证了算法的可行性。

Rough Set Theory is another mathematical tool used for dealing with fuzzy and uncertain knowledge besides Probability Theory, Fuzzy Set Theory and Evidence Theory. In order to solve the problems with the same condition attributes but different decision - making attributes, which often appear in the use of Rough Set Theo- ry, we put forward an algorithm of correcting incompatible rules. In other words, according to the concept of relative error rates of classification, under the precondition of maintaining the original knowledge classification unaltered, we correct the incompatible rules though comparison of the magnitudes of the relative error rates of classification. At last, we verify the feasibility of the arithmetic with an experience of classification of radars.