摘要： 在优势信息系统中引入粒度熵概念，定义其相关的性质和定理，并以优势关系下的粒度熵为启发式信息，给出属性的重要性度量。在此基础上提出一种基于粒度熵的启发式属性约简算法，并分析该算法的时间复杂度。通过实例说明该算法的有效性，为进一步研究优势目标信息系统的规则提取提供理论基础。

关键词: 粗糙集, 优势关系, 粒度熵, 属性约简, 启发式算法

Abstract: The concept of granular entropy is introduced in dominance information systems, the granular entropy based on dominance relation is defined, its related properties and theorems are gained, the significance of attribute is defined through granular entropy, with the significance of attribute as heuristic information. On this basis, a heuristic algorithm based on granular entropy for reduction of attribute is proposed, and the time complexity of the algorithm is analyzed. It illustrates that this algorithm is effective through an example, which puts forward new theoretical basis for further researching of rule extraction in target information systems based on dominance relation.

Key words: rough set, dominance relation, granular entropy, attribute reduction, heuristic algorithm

中图分类号: 

• TP18