一种基于粗糙熵的信息系统属性约简算法

An Attribute Reduction Algorithm Based on Rough Entropy in Information System

在信息系统中,研究了知识的粗糙性,定义了一种粗糙熵度量方法,并证明了知识的粗糙熵随着划分的增大而单调增加的结论,给出了属性的重要性度量方法,在此基础上提出了一种基于粗糙熵的启发式属性约简算法。实例验证表明,该算法能有效地从信息系统中获取最优属性约简。

In information system,a rough entropy is defined by studying roughness of knowledge,then knowledge rough entropy’s mo notonous increasing property with the increase of partition is proved.On this basis,attribute significance measure method is given and a heuristic reduction algorithm based on rough entropy is proposed.A detailed example is shown that the algorithm can effectively extract op timal reduction.