Abstract: In order to process incomplete information systems，the classical rough set theory needs to be extended.In this paper，the incomplete information system is investigated.In such information systems，a new relation is proposed which is based on one parameter，and the rough set model is based on this new relation.An extended model is obtained，which is more effective in classification if setting the parameter properly.It is proved that the new model can anticipate and classify data more precisely.A practical example has verified the advantage of the new binary relation and the extended model.

Key words: rough set, Incomplete Information System（IIS）, tolerance relation, non-symmetrical similarity relation, limited tolerance relation, improved binary relation

摘要： 对于不完备信息系统的处理，需要对经典粗集理论进行扩充。以不完备信息系统为研究对象，提出了一种新的带有参数的二元关系，并讨论了基于该二元关系的粗糙集扩充模型。通过合理设置参数值，可使新扩充模型比以往模型具有更好的分类效果，从而提高对数据预测、分类的准确度。实例分析进一步验证了新二元关系及其扩充模型的优越性。

关键词: 粗糙集, 不完备信息系统, 容差关系, 非对称相似关系, 限制容差关系, 改进的二元关系