基于本体的贝叶斯网络知识推理研究  

Research of Bayesian Network Reasoning Based on Ontology

作　　者：蔡炳万 石宇强[1] 李明辉[1] 张敏[1] 

机构地区：[1]西南科技大学制造科学与工程学院,四川绵阳621000

出　　处：《机械设计与制造》2016年第1期84-87,共4页Machinery Design & Manufacture

摘　　要：针对产品设计过程中因设计知识不断增多、知识存储位置分散且知识表示没有统一规范而导致的知识利用效率低下、知识推理精度不高的问题,提出了一种基于本体的贝叶斯网络知识推理的方法。以包装产品设计过程为例,根据整个设计过程中的知识活动,建立了包装设计的设计任务、设计知识、设计人员的知识本体模型;依据本体模型中描述的概念—概念间的相互关系,建立网络拓扑结构,定性的描述节点间的关系,并根据历史的设计方案以及专家经验确定网络节点间条件概率分布。最后,根据贝叶斯理论实现知识的推理。Due to the increasing of design knowledge during the product design process,knowledge is stored in scattered locations and the knowledge express without unified standard,leading to inefficient use of knowledge,knowledge inference precision being low,against above problem,it proposed Bayesian network knowledge reasoning based on ontology. Taking packing design as an example,according to the knowledge activity of packaging products design process,it establishedthe packaging design tasks,design knowledge,knowledge ontology model designers; Based on ontology model described- the relationship between concepts,established Bayesian network topology,anddetermined the quantitative relationship between network nodes based on expert experiences and historical design. Finally,according to the theory of Bayesian network inference ontology domain knowledge to achieve.

关 键 词：本体 贝叶斯网络 知识推理 包装设计 

分 类 号：TH16[机械工程—机械制造及自动化]