基于字典的半自动常识本体构建

摘要

To ensure the stucture and semantic of the exchanged information are consistent, this paper proposed a semi-automatic common ontology construction method based on dictionary in order to construct unified standard common-sense ontology. This method took the dictionary as the source of ontology, and combined the ontology expansion mechanism based on free text, it could constuct and expand ontology semi-automatically relying on extracting concept' s attribute relationship from its interpretation. Experiments show that the method constucts ontology semi-automatically, and' the common ontology contains more than 70 000 common-sense concept, 600 000 assertions and 1 900 predicates.%为了保证所交换的信息在结构和语义上的一致性,实现真正意义上的信息互操作与集成,提出一种基于字典的半自动本体构建方法,并运用该方法构建一个具有统一标准结构的常识本体.基于字典的本体构建方法将字典给出的常识性知识作为本体的来源,并结合基于自由文本的本体扩展机制,从概念的解释中抽取其属性关系,半自动地构建和扩展常识本体.实验表明,该方法能够在自动化程度较高的基础上,构建一个包含7万多个常识性概念、60万个断言和1 900个谓词的常识本体.