摘要： 基于不协调优势目标信息系统，给出刻画属性重要性的新指标，以该指标为启发式信息，提出一种新的下近似约简算法。该算法具有较大的灵活性，能从搜索空间逐次删除不重要属性，避免重复计算属性的重要性，从而提高搜索的效率，且时间复杂度较低。通过实例说明了该算法的有效性。

关键词: 粗糙集, 信息系统, 优势关系, 下近似约简, 启发式算法

Abstract: This paper puts forward a new relatively reasonable formula measuring attribute significance based on inconsistent dominance object information systems. On this basis, a new heuristic algorithm based on lower approximation for reduction of attribute is proposed. Compared with those existed algorithms, its flexibility is increased because calculating the important value of unimportant attributes repeatedly can be avoided by removing unimportant attributes gradually from the search space, and this algorithm is much less time complexity than those existed algorithms. Example illustrates the effectiveness of this algorithm.

Key words: rough set, information systems, dominance relation, lower approximation reduction, heuristic algorithm

中图分类号: 

• TP18