论文部分内容阅读
文中在粗糙集理论中的约简概念的启发下提出了极小规则和极大规则的概念及极小极大规则学习．它可用于简化决策树归纳学习方法得到的规则．决策树方法是当前数据库知识发现中所采用的最有效的一种分类规则学习方法，因此本方法具有潜在的实用意义．实验结果表明采用极小极大规则学习简化决策树规则，既能简化单个规则，又能减少规则总的数量．
Inspired by the concept of reduction in rough set theory, the concepts of minimal rules and maximal rules and minimal maximal rules learning are proposed. It can be used to simplify the rules obtained by the decision tree induction learning method. The decision tree method is the most effective classification rule learning method used in the current database knowledge discovery, so the method has potential practical significance. The experimental results show that using minimal maximal rule learning to simplify the decision tree rules can not only simplify a single rule but also reduce the total number of rules.