基于属性加权频率算子的属性约简算法

Attributes Reduction Algorithm Based on Attributes Weight Frequency Function

定义了属性加权频率算子的概念,并在此基础上提出一种新的属性约简算法,理论分析和实验结果表明,该算法可以确保得到决策表的一个约简,并能提高算法的效率。同时引入强等价集,很好地解决了属性加权频率值相同的问题。

This paper defines Attributes Weight Frequency Function（AWFF） and proposes a new attributes reduction algorithm. Theoretical analysis and experimental results show that the algorithm can get a reduction and improve the algorithm efficiency. Meanwhile the strong compressible set has been introduced, to solve the problem that the values of attributes weight frequency are the same.