首页|雷达知识图谱构建方法及应用

雷达知识图谱构建方法及应用

扫码查看
描述雷达及其属性的非结构化文本信息分散在互联网各处,信息量丰富但数据格式多样,难以直接处理和分析,构建雷达知识图谱能够有效解决这一问题.为此本文首先梳理了构建雷达知识图谱的流程;然后结合雷达领域的特点同时参考七步法,构建了雷达知识图谱的本体层;最后利用互联网文本语料作为数据源,对通用信息抽取(UIE)模型进行微调,完成了知识抽取,构建了雷达知识图谱.应用结果表明,构建的雷达知识图谱可为业内人员提供参考查询的功能,对搭载雷达的战斗平台进行高维画像;同时可作为底层数据库进一步开发,支撑完成上层应用.
Construction method and application of radar knowledge graph
The unstructured textual information describing radar and its attributes is scattered all over the In-ternet.The information is abundant,but the data formats are diverse,so it is difficult to directly process and ana-lyze them.Constructing radar knowledge graph helps effectively solve the problem.Therefore,this paper first sorts out the process of constructing radar knowledge graph,and then combines with the characteristics of radar domain and simultaneously refers to the seven-step method to construct the ontology layer of radar knowledge graph.Finally,the paper uses the Internet text corpus as data source to fine-tune the universal information extrac-tion(UIE)model,with the knowledge extraction completed,and the radar knowledge graph constructed.The ap-plication results show that the constructed radar knowledge graph can provide the function of reference query for the industry personnel and high-dimensional portrait of the combat platform equipped with radar,and meanwhile,can be further developed as the underlying database to support the completion of the upper layer application.

radar knowledge graphontology layer of radar domainknowledge extractionuniversal infor-mation extraction(UIE)

黄振铭、吴晓芳、薛孟武

展开 >

军事科学院系统工程研究院,北京 100191

雷达知识图谱 雷达本体层 知识抽取 通用信息抽取

2024

空天预警研究学报
空军预警学院

空天预警研究学报

影响因子:0.39
ISSN:2097-180X
年,卷(期):2024.38(3)
  • 2