变论域知识约简算法

从论域的角度出发,提出了一种变域论域知识的简算法。该算法需要多次计算属性重要度，在计算过程中，不但论域逐步减少，有助于减少计算量,而且相对属性集不断增加，有助于选择出反映特征的重要属性来。实验证明，该算法不失为一种较好的知识约简算法。

Variable Universal Knowledge Reduction Algorithm

From the viewpoint of the universal this paper proposed an algorithm which is called the variable universal reduction algorithm. The algorithm needs to compute the significance of the attributes many times. During computing, the universal is decreasing, which helps to decrease the computing complexity, and the relative attributes set is increasing, which conduces to select the attributes that reflect the characteristic of the system. The experimental results show that this algorithm can usually find the better relative reduction of a decision system.