基于可辨识矩阵的属性频率约简算法

An algorithm of attribute frequency reduction based on discernibility matrix

阐述了粗糙集理论的基本概念,知识约简是粗糙集理论研究的核心内容之一,将可辨识矩阵中出现次数多少作为属性重要性的判断依据．算法求取的是所有约简结果中的某一个或某一些(当某两个条件属性出现频率相同时)结果,当信息系统的复杂程度较高时,其求解的复杂度大大小于原来的约简方法．

This paper expounds the basic conceptions of the rough sets theory, in which knowledge reduction is a very important topic. An algorithm which is based on the frequency of attribute is proposed in the discernibility matrix of an information system. The algorithm produces some result of all attribute reductions （when the attributes of two conditions appear at the same the frequency）. Using the algorithm we can reduce the complexity of the original attribute reduction algorithm greatly and acquire the optimal decision-making rules of the information system.