模糊信息表决策规则获取与属性约简方法

将粗糙集方法用于模糊信息表上的规则获取和属性约简, 在包含度概念的基础上, 定义了齐次包含度.对两类模糊信息值的信息表, 条件属性是离散值，而决策属性是模糊的模糊决策信息表和条件属性与决策属性均为模糊的模糊信息表, 采用齐次包含度研究了条件属性在决策属性中的包含关系, 得到了包含度意义下模糊决策信息表的属性约简判定定理和决策规则获取方法. 给出了模糊条件属性的所有可能组合, 研究了模糊组合属性在模糊决策属性中的包含关系, 提出了最大属性协调集的概念，建立了在最大协调属性集下的模糊信息表属性约简和决策规则获取的算法步骤, 并通过算例演示了模糊决策规则获取和属性约简的过程, 从而为模糊信息表决策规则获取与属性约简提供了一类算法.

Approach to decision rule acquisition and attribute reduction on fuzzy information table

Rough set theory is applied for rule acquisitions and attribute reductions on fuzzy information tables. Based on the notion of inclusion degrees, a kind of homogeneous inclusion degrees is defined. Two types of information tables with fuzzy numbers, called the fuzzy decision information table and the fuzzy information table in which only decision attributes are fuzzy and both condition and decision attributes are fuzzy respectively, are discussed. Relations of condition attributes including decision attributes are studied by inclusion degrees and the judgment theorem of attribute reductions and the approach to decision rule acquisitions are obtained under the meaning of inclusion degrees. All possible compositions of fuzzy condition attributes are given, relations of the fuzzy composed attributes including fuzzy decision attributes are discussed and the notion of maximal consistent attribute sets is introduced. The algorithm step of attribute reductions and decision rule acquisitions is established under the meaning of maximal consistent attribute sets on fuzzy information tables. Some examples are illustrated and the process of fuzzy decision rule acquisitions and attribute reductions is explained. Thus a kind of algorithms is supplied for decision rule acquisitions and attribute reductions on fuzzy information tables.