Abstract: This paper proposed a domain ontology construction method based on knowledge graphs. This domain ontology was constructed based on the mature idea of software engineering by using the "HowNet" as the semantic knowledge resources, and knowledge graphs as a semantic representation. It has a clear structure with clear semantic meaning. It can provide effective supports for some applications such as semantic Web and information extraction. This paper introduced the concept of ontology, the standard of design, the process of modeling, and future prospects of ontology portability. This method was proved to be more effective than traditional ontology construction method in uncertain knowledge processing by experimental results.

Key words: ontology, domain ontology, semantic, knowledge graph

摘要： 提出了一种基于知识图的领域本体半自动构建方法。以《知网》为语义知识资源，知识图为语义表示方法，采用成熟的软件工程流程，最终构建出的领域本体具有结构明确、语义清晰的特点。对于在其上的语义网、信息抽取等应用提供了有效支持。介绍了本体的概念、设计的准则、建模的流程，并对未来的本体的移植性进行展望。实验结果表明该方法在不确定性知识处理上优于传统本体构建方法。

关键词: 本体, 领域本体, 语义, 知识图