摘要： 根据优势原理，提出一种具有明确粗糙集理论含义的指标——优势度，用于度量序目标信息系统的协调程度。在证明优势度粒化单调性的基础上，给出属性集重要性度量函数，提出一种基于优势度的序目标信息系统启发式约简算法。该算法与经典粗糙集理论约简有相同的理论基础，易于理解。应用结果表明，该算法适用于优势关系下目标信息系统的知识发现。
关键词:
粗糙集,
属性约简,
优势关系,
不协调信息系统,
属性重要性
Abstract: A new uncertainty measure, such as dominance degree is proposed in ordered objective information systems based on dominance principle, and an explicit theoretical meaning of rough set is given to the dominance degree which can be used to measure the inconsistence of objective information system. The granulation monotonicity of dominance degree is proved, based on which a new measure of attribution importance is designed. An heuristic reduct algorithm in objective information system is provided based on dominance relation. An example illustrates the validity of this algorithm, and results show that the algorithm has the same theoretical foundation with classical reduct algorithm in rough set theory, and it is easily understood. The algorithm provides an important theoretical basis for knowledge discovery in ordered objective information systems.
Key words:
rough set,
attribute reduction,
dominance relation,
inconsistent information system,
attribute importance
中图分类号:
廖帆, 滕书华, 邵世雷. 基于优势关系的启发式属性约简算法[J]. 计算机工程, 2011, 37(24): 52-54.
LIAO Fan, TENG Shu-Hua, SHAO Shi-Lei. Heuristic Algorithm for Attribute Reduction Based on Dominance Relation[J]. Computer Engineering, 2011, 37(24): 52-54.