基于多源异构数据的甲骨学知识图谱构建方法研究

Oracle bone studies knowledge graph construction based on multi-source heterogeneous data

为解决和缓解甲骨学研究难度大、周期长、知识关联性弱、知识共享程度低等问题。基于多源异构数据源,融合基于文献计量学的科学知识图谱(MKD)和基于知识库的知识图谱(KG),构建了甲骨学融合知识图谱。通过融合两类知识图谱,并基于知识推理进行语义扩展,形成最终的甲骨学知识图谱。其中包含实体148 305个,关系434032条,可满足甲骨学研究的基本需求。融合MKD和KG两类知识图谱,优势互补,实现甲骨学知识图谱构建,可为其他古籍类知识图谱构建提供借鉴。

In order to solve or alleviate the problems such as great difficulty, long period, weak knowledge linking and low sharing regarding the oracle bone studies, we construct the oracle bone studies knowledge graph by combining two kinds of knowledge graphs MKD and KG.These two kinds of knowledge graphs are fused and semantic extension is carried out based on knowledge reasoning. The constructed oracle bone studies knowledge graph contains 148 305 entities and 434 032 relations, which can meet the basic needs of oracle bone studies research. It also provides reference for other ancient books knowledge graph construction.