经典粗集理论以等价关系为基础,它们在处理不完备信息系统的时候存在一定的局限性.作者首先从扩展粗集理论适应范围的角度出发,引入信息系统对象的先验概率,结合量化相容关系的思想,提出基于先验概率的相容关系模型,然后以该模型为基础研究了不完备信息系统的属性约简,通过严格的数学证明,得到基于分辨矩阵的属性约简算法.最后通过一个实例验证了该改进相容关系模型及其基于分辨矩阵的约简算法.表1,参20. Classical rough set theory is based on equivalence relations, and they have some limitations when dealing with incomplete information system.First, from the perspective of extending the range of rough set theory, the author introduces the prior probability of information system objects, Compatibility relation, this paper puts forward a conformable relation model based on prior probability, and then studies the attribute reduction of incomplete information system based on this model. Through rigorous mathematical proofs, the attribute reduction algorithm based on discernibility matrix is ​​obtained. Finally, an example is given to verify the improved compatibility model and its resolution matrix-based reduction algorithm.