论文部分内容阅读
为了克服目前在基于规则推理(Rule-Based Reasoning,RBR)系统中获取和更新规则的困难以及加快基于事例推理(Case-Based Reasoning,CBR)系统中的事例检索过程,文章提出了一种基于模糊集和粗糙集理论、从冲压模具设计库中发掘隐性设计知识的技术。所发掘的知识以规则的形式表达。该知识发掘过程包括:(1)冲压零件基于特征的模糊表达;(2)冲压事例的模糊聚类;(3)基于粗糙集的规则获取。为了进行验证,文章在原型系统中建立了一个包含53个事例的原型事例库,并从中自动发掘出了38条规则。通过事例聚类和属性约减,CBR系统的检索速度可以得到较大的提高。此外,由于所发掘的规则以产生式规则的形式表达,RBR系统可以利用这种技术建立和更新规则。
In order to overcome the current difficulties in obtaining and updating rules in Rule-Based Reasoning (RBR) systems and to speed up the case retrieval process in Case-Based Reasoning (CBR) systems, this paper proposes a fuzzy- Set and rough set theory, from the stamping die design library to explore the technology of hidden design knowledge. The knowledge that is excavated is expressed in the form of rules. The knowledge mining process includes: (1) feature-based fuzzy expression of stamping parts; (2) fuzzy clustering of stamping cases; and (3) rule-based rough set acquisition. For validation, the article built a prototype case library of 53 cases in the prototype system and automatically excavated 38 rules. By case clustering and attribute reduction, the CBR system retrieval speed can be greatly improved. In addition, RBR systems can use this technique to establish and update rules because the rules they discover are expressed in terms of generative rules.