摘 要：针对工程机械维修保养知识图谱建模人工建模成本高、难度大且覆盖度难以保证等问题，提出基于自顶向下和自底向上相结合的知识图谱建模方法。首先，人工定义构建工程机械维修保养知识图谱的顶层数据模式；然后，利用知识抽取技术从文本数据中抽取实体、属性、关系知识；最后对自顶向下方法的结果与自底向上方法的结果进行合并融合。实验验证表明，该文所提出的知识建模方法可以大大降低人工建模的难度，同时具备较好的覆盖度和准确性，可以满足工程机械维修保养知识图谱构建与应用的需求。
关键词：知识体系；动态本体；工程机械维修保养；知识图谱；自然语言处理
中图分类号：TP391 文献标识码：A 文章编号：2096-4706（2020）07-0011-04
Research on the Knowledge Map Modeling of Construction Machinery Maintenance
DONG Dengkui，ZENG Yonghua
（Army Engineering University，Nanjing 210007，China）
Abstract：Aiming at the problems of high cost，difficulty and difficulty in guaranteeing the construction of knowledge graph modeling for construction machinery maintenance knowledge，a knowledge graph modeling method based on the combination of top-down and bottom-up is proposed. First，manually define the top-level data model for constructing the knowledge map of maintenance and repair of construction machinery；then，use knowledge extraction techniques to extract entity，attribute，and relationship knowledge from the text data；finally，the results of top-down method and bottom-up method are merged. Experimental verification shows that the knowledge modeling method proposed in this article can greatly reduce the difficulty of manual modeling，and has good coverage and accuracy，which can meet the needs of construction machinery maintenance knowledge graph construction and application.
Keywords：knowledge system；dynamic ontology；maintenance of construction machinery；knowledge graph；natural language processing
参考文献：
[1] AMIT S. Introducing the knowledge graph [R].America：Official Blog of Google，2012.
[2] 车金立，唐力伟，邓士杰，等 . 装备维修保障知识图谱构建方法研究 [J]. 兵工自动化，2019，38（1）：15-19.
作者简介：董登奎（1987.06—），男，汉族，安徽滁州人，硕士在读，研究方向：工程机械维修保养。