一种新的决策表属性值分类方法

A new classification method for attribute values in a decision table

在粗糙集值约简算法中,常常需要对决策表的条件属性值进行分类。基于Skowron分辨矩阵,提出一种新的属性值分类矩阵。通过该矩阵可以方便的获取决策表中各条件属性的冲突记录集和重复记录集,并能够确保对这些冲突记录集和重复记录集的条件属性值进行处理后,决策表中剩下的条件属性值被删除后不会产生新的冲突记录和重复记录。理论分析与实例表明:该方法能有效处理决策表属性值分类。

In rough set attribute value reduction, it is necessary to classify condition attribute values in a decision table. A new classification matrix for attribute values is put forward based on Skowron discernible matrix. By using this matrix, we can easily obtain the conflict record set and duplicate record set of each condition attribute in a decision table, and no more other conflict records or duplicate records can be produced after handling the corresponding condition attribute values of these conflict record sets and duplicate record sets and deleting the rest of condition attribute values in a decision table. The theoretical analysis and example show that this method can effectively deal with attribute value classification in a decision table.