In view of the limitations of the attribute reduction algorithm based on variable precision rough set in tolerance rela-tion, the judgment theorems and discernibility matrixes with respect to β upper(lower)distribution reductions are presented.The method for β upper(lower)distribution reduction in incomplete decision table is also presented. At last, the experiment demonstrates the effectiveness of the proposed method.%针对现有的在相容关系下基于变精度粗糙集约简算法的局限,给出了β上(下)分布约简的判定方法和改进的β上(下)分布可辨识矩阵定义,并给出求解基于变精度粗糙集的不完备决策表约简算法;最后通过实例验证了算法的有效性。
展开▼