摘要
提出一种基于Skowron分辨矩阵的既能学习决策分类规则又能学习非决策分类规则方法,目的是获得更一般、更可靠的分类规则,并设计了算法.算法的基本思想是以Skowron分辨矩阵讨论非一致决策系统的属性约简的结果为基础,获得基于分辨矩阵的非一致决策系统的确定性规则.应用例子表明,所提出的算法获得的规则为简洁和高效.
A method of rule induction based on Skowron's discemable matrix,which is capable of learning both deterministic and non-deterministic classification rules,is discussed.The method makes the classification rules more general and reliable.An algorithm is put forward.The basic idea is that basing on Skowron's discemable matrix researching attribute reduction of inconsistent decision system,acquisition deterministic rules of inconsistent decision system with Skowron's discemable matrix as a base. Examples show that the algorithm provides more precise and simple decision rules.
出处
《三明学院学报》
2006年第2期194-197,共4页
Journal of Sanming University
关键词
粗集
信息系统
分辨矩阵
决策规则
Rough Set
Information System
Discemable matrix
Decision Rule