基于粗集理论的决策表知识获取方法研究

Research on knowledge acquisition from decision table based on rough sets theory

对决策表各条件分类和决策分类集合之间的关系进行了研究,提出了直接从各分类中计算决策表核及属性约简方法;依据属性约简,创建了一个多变量决策树;在此基础上,阐述了提取决策表中蕴含规则的方法,从而省去了在约简后的决策表中计算值约简步骤;通过实例,验证了这些方法的有效性。

The relationship between conditional equivalence classes and decision equivalence classes are researched in detail. From those classes, methods of core and attribute reduction are presented, According to attribute reduction, a similarity decision tree is created. On the basis of the tree, a method of acquisition decision rules is presented, and the advantage of using this method is to omit value reduction in decision table, An example illustrates the efficiency of those methods.