Abstract: According to the definition and characteristic of incomplete information system, this paper establishes rough set model based on set error tolerance relationship, constructs discernable matrix of condition attribute, and analyzes importance of decision attribute. It expatiates and compares three kinds of attribute reduction method. Processing results on one incomplete information table show that the three kinds of method have the almost same reduction results and validity.

Key words: incomplete information system, rough set, attribute reduction

摘要： 根据不完备信息系统的定义和特点，建立基于集合容差关系的粗糙集模型，构造条件属性的可辨识矩阵，分析决策属性的重要性。阐述并比较3种属性约简方法，对同一个不完备信息表的处理结果表明，3种方法可以得到基本一致的约简结果，验证了其有效性。

关键词: 不完备信息系统, 粗糙集, 属性约简

CLC Number: 

• TP391