属性约简是粗糙集理论重要研究内容之一,基于可分辨矩阵的属性约简方法需占用大量存储空间,不利于大数据集的处理.为此,引入差别集定义和基于差别集属性约简定义,并指出基于差别集属性约简本质上是在当前差别集中不断寻求关键属性的过程,并给出删除单个条件属性和删除条件属性集两种获取关键属性的属性约简方法,同时证明了这两种属性约简方法是正确的、完备的;进一步,为了获得最小属性约简,采用两个启发式信息来筛选关键属性;在上述基础上,设计基于差别集的启发式属性约简算法.最后,通过实例和实验验证了该算法的有效性和高效性. Attribute reduction is one of the most important research topics in Rough Set Theory. The attribute reduction method based on Distinguishing Matrix requires a large amount of storage space and is not conducive to the processing of large data sets.Therefore, the definition of differential set and the attribute set of difference set It is pointed out that attribute reduction based on difference sets is essentially the process of seeking key attributes in the current difference set. Two attribute reduction methods are given to get the key attributes, which are to delete the single condition attributes and the deletion condition attribute sets. At the same time, it is proved that The two attribute reduction methods are correct and complete. Furthermore, in order to obtain the minimum attribute reduction, two heuristic information are used to filter the key attributes. Based on the above, the heuristic attribute reduction algorithm based on the difference set is designed Finally, the validity and efficiency of the algorithm are verified by examples and experiments.