在序信息系统中,研究了知识的粗糙性与知识划分的关系,为准确度量知识的粗糙程度,定义了知识的粗糙熵和属性的重要性度量,并给出了相关的性质。为有效地从序信息系统中获取最小属性约简,基于知识的粗糙熵提出了一种新的启发式属性约简算法,并通过实例验证了该方法的有效性。
In order information systcm, relation bctween knowledge rough and knowledge partition are studied. To accurately measure knowledge rough, knowledge rough entropy and attribute significance are defined, and then related properties are given. A heuristic reduction algorithm based on knowledge rough entropy is proposed to effectively extract minimum attribute reduction and a detailed example is given to prove validity of the algorithm.