在基于等价关系的随机信息系统中,文章以证据理论中的信任测度和似然测度为基本工具,给出了核心属性、不必要属性及相对必要属性的一些等价刻画,研究了随机目标信息系统的属性约简问题,并利用实例说明了约简方法的有效性。
By means of the belief measure and the plausibility measure in the evidence theory, the e- quivalent characterization of the importance of core attribute, unnecessary attribute and relatively nec- essary attribute in the equivalent relation based random information systems is discussed. And then the attribute reduction in the random objective systems is studied. Finally, an example is used to illus- trate the validity of this method.