一种改进的基于区分矩阵的属性约简算法

Improved Attribute Reduction Algorithm Based on Discernibility Matrix

根据粗糙集理论讨论了属性约简问题,基于传统的区分矩阵属性约简算法提出了一种改进的约简算法:首先,根据决策表构造区分矩阵;然后,删除区分矩阵中包含核和伪核的项;最后,化简区分矩阵中剩余项得到约简结果.通过对某发动机故障数据进行处理验证了算法的有效性.

This paper discusses the problem of attribute reduction according to rough set theory, and proposes one improved attribute reduction algorithm based on the traditional discernibility matrix. First, the discemibility matrix is constructed according to the decision table. Then, the items containing core and fake core are deleted from the discernibility matrix. Finally, the reduction result is obtained by simplifying the remaining items of the discernibility matrix. This paper verifies the effectiveness of the algorithm by processing engine failure data.