Improved Attribute Reduction Algorithm Based on Discernibility Matrix

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 discernibility 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.