一种专家系统知识获取时的属性值约简算法

An Attribute Value Reduction Algorithm of Expert System Knowledge Acquisition

知识获取是构造专家系统的"瓶颈",提供准确的推理知识是进行科学决策的关键。文中运用粗糙集理论,研究对决策表中每条记录的冗余条件属性值进行筛选并删除的属性值约简算法。首先研究属性值约简的理论基础,包括知识表示和知识约简与核两个方面;其次研究知识获取方式与知识获取过程;然后研究属性值约简算法,通过两个定义描述约简算法的基础上,给出了约简算法的5个步骤;最后以城市物流中心选址为例,运用属性值约简算法及其步骤,对决策表属性值进行约简。结果表明,属性值约简实现了决策表的最简化,突出了关键属性及其关键属性值对决策的影响。

To construct an expert system, knowledge acquisition is the ＂bottleneck＂ problem, and the key to scientific decision-making is accurate inference knowledge. In this paper, it will use rough sets theory to attribute value reduction algorithms that filter and remove the redundant condition attribute value of decision-making table＇s each record. First,research the theoretical basis of attribute value reduction, including knowledge representation and knowledge reduction and core both. Then, research knowledge acquisition ways and knowledge acquisition processes. Third, research the attribute value reduction algorithms, through the description of two definitions the reduction algorithms is given the 5 steps. Finally, take the city logistics center location as example, use attribute value reduction algorithms and its steps for the value of decision table attribute reduction. The results show that attribute value reduction achieves the most streamlined decision-making table, highlighting the key attributes and key decision-making of property value.