Abstract:
Taking notice that condition attribute whose value equals with “*” does not participate in the determination of similarity relation in the incomplete decision table based on similarity relation, this paper defines relative important degree and important degree of attributes. According to the important degree of attributes, the heuristic method of searching for assignment consistent sets and assignment reduction in incomplete decision table are presented respectively. Experimental results show that the method can find the assignment consistent sets and assignment reduction of an incomplete decision table, and reduce the search space of reduction.
Key words:
incomplete decision table,
similarity relation,
rough set,
attributes reduction,
assignment reduction
摘要: 在基于相似关系的不完备决策表中,条件属性值为“*”的属性不参与相似关系的确定,由此定义属性的相对重要度和属性重要度。根据属性重要度,提出寻找不完备决策表的分配协调集和分配约简的启发式算法。通过实例证明该方法能找到不完备决策表的分配协调集和分配约简,并减少约简的搜索空间。
关键词:
不完备决策表,
相似关系,
粗糙集,
属性约简,
分配约简
CLC Number:
YANG Ji-Lin, QIN Ke-Yun, FEI Zheng. Attributes Reduction Based on Similarity Relation in Incomplete Decision Table[J]. Computer Engineering, 2010, 36(20): 10-12.
杨霁琳, 秦克云, 裴峥. 不完备决策表中基于相似关系的属性约简[J]. 计算机工程, 2010, 36(20): 10-12.