Knowledge reduction algorithm based on elementary row transformation of Boolean matrix
基于布尔矩阵的初等行变换的知识约简算法

The definition of elementary row transformation of Boolean matrix was given. Attribute reduction model in the form of linear logical equation group was established. Coefficient matrix was transformed into simplest matrix by elementary row transformation of Boolean matrix. Three necessary and sufficient conditions to determine whether the attribute was absolutely necessary, relatively necessary and absolutely superfluous were given by coefficient matrix and simplest matrix, and a fast knowledge reduction algorithm was proposed.