基于属性关系积的值约简算法
首发时间:2009-03-14
摘要:属性值约简是粗集理论研究的重要内容之一。本文将粗糙集理论应用于一致决策表和不一致决策表的值约简,给出了属性关系积、规则的准确度等定义和约简规则的判定准则,在此基础上提出了一种基于属性关系积新的值约简算法。算法采用逐步扩大属性子集的宽度优先搜索策略,把对决策表的值约简过程转化成属性关系积的运算,规则产生的指导思想是:对于一致规则考察其在决策表中约简前后是否一致;对于不一致规则考察其在决策表中约简前后准确度是否发生变化。算法只需对决策表扫描一次,提高了属性值约简的效率,最后通过对该算法进行描述和实例分析验证讨论了算法的有效性。
For information in English, please click here
A Value Reduction Algorithm Based on Attribute Union
Abstract:Attribute value reduction is one of the important parts researched in rough set. This paper applies rough set theory to the consistent decision table and inconsistent decision table of the value reduction. It gives the definitions of attribute union, the measure standard of the accuracy of rules, and discusses the determination methods to extract rules. According to these methods, this paper proposes a new value reduction algorithm based on attribute union. The algorithm use the scanning strategy with width priority of gradually expand the subset property, It convert the attribute reduction to attribute union computing,the guiding ideology of rule extraction is that consistent rules maintain consistency and inconsistent rules maintain accuracy of rules in decision table.The algorithm scan the decision table only one time and improve the reduced efficiency. Finally, the description of the algorithm and analysis of the experiments verify the effectiveness of the algorithm.
Keywords: Rough set value reduction attribute union accuracy of rules
基金:
论文图表:
引用
No.3026043735512370****
同行评议
共计0人参与
勘误表
基于属性关系积的值约简算法
评论
全部评论0/1000