利用粗集原理完善评估指标体系的方法

Refining assessing system using rough set theory

研究在建立评估指标体系的过程中 ,利用粗集理论检验和挑选重要指标集的方法 .该方法利用熵函数和MDL原理设计出评价指标集与最后评分间因果关系的度量函数 ,通过分析指标子集与最后评分结果之间的关系 ,挑选出最有效的指标子集作为最终的评估依据 .利用该方法 ,可以去除指标系统中的无关指标 ,在不降低评估精度的前提下 。

In this paper,we studied the method of using rough set theory to select the most useful features in assessing system.The method uses the concept of entropy and MDLP to define the evaluation function of a feature subset.And then,we use this function to search the best feature subset among all features in the assessing system.Using the algorithm in this paper,we can get a smaller feature subest with the same assessing result of using all features.Thus we can reduce the assessing system.