摘要
Rough集理论为知识约简提供了一种有效的方法.提出了一种基于遗传算法的Rough集多知识抽取方法.针对决策系统中知识约简的不唯一性,构造了一种多约简算法,创建了多知识.在此基础上,利用遗传算法从一个更高的层次对多知识进行优化,并从中抽取最优知识集.试验结果分析表明,通过遗传算法优化后抽取的多知识较单体知识具有更高的精度,使知识的表示更具广义性.
Rough theory provides an effective approach for knowledge reduces. A genetic algorithm based multi-knowledge extraction method is proposed for rough set. Multi-reducts algorithm is constructed and multi-knowledge created according to much knowledge reducts in decision system. Based on the above result, the multi-knowledge is optimized by genetic algorithm from a more high level, and the optimized knowledge is extracted. Comparing with single body knowledge, experimental analysis and results show that the multi-knowledge optimized and extracted enhance the precision and can make the knowledge representation more generalizable.
出处
《小型微型计算机系统》
CSCD
北大核心
2005年第4期651-654,共4页
Journal of Chinese Computer Systems
关键词
粗糙集
多知识
遗传算法
知识约简
rough set
multi-knowledge
genetic algorithm
knowledge reducts
