决策分析中属性约简的择优算法

粗糙集理论在决策分析中具有广泛的应用。基于优势关系的粗糙集理论只给出如何进行属性约简以得到决策规则的方法，本文则进一步研究了如何针对不同的属性约简如何进行选择。在优势关系的基础上引入格序的概念，利用序关系给出属性约简的贴近度，借此比较了不同约简下所得决策规则贴近于原知识库的程度，并通过一个具体例子加以说明。

An Algorithm to Choose the Reduction of Attributes in Decision Analysis

The theory of the rough set is widely used in decision analysis. The theory of rough set based on dominance relation only presents us with how to obtain the decision rules by attributes reductions, while this thesis gives a further study on how to make a choice according to different attribute reduction. After the definition of lattice order based on dominance relation, the fidelity of a reduction is defined through order relation, through which a comparison of close degree is made between the decision rules obtained by different reductions and original table. In the end, an illustrative example is given to show the algorithm exactly.