基于关联矩阵的属性约简算法

An Attribute Reduction Algorithm Based on Association Matrix

利用差别矩阵对信息系统特别是对大规模数据的信息系统进行属性约简研究的一个重要方面就是如何提高计算速度。为改进差别矩阵的应用,提高约简效率,文章提出了关联矩阵的概念,同时,依照决策属性对条件属性的依赖程度,利用关联阵中属性频率的信息,提出了一种属性约简算法。实验结果证明了该算法有效、快捷。

Attribute reduction is one of the key problems for the rough set theory.It is important how to get a reduction of an information system more efficiently.Based on the concept of discernibility matrix,a new concept called association matrix is proposed in this paper,and an attribute reduction algorithm is given accordingly.Experimental results show the algorithm is fast and effective.