古汉语典籍承载着中华优秀传统文化，典籍引书是发现传统文化传承脉络的重要指南，而知识图谱构建则代表了人工智能时代的技术前沿。本项目以“基于深度学习的典籍引书知识图谱构建及应用研究”为题，以具有代表性的古汉语数字化典籍为数据来源，在文本预处理和人工校勘以及浅层句法分析的基础上构建典籍语料库。研究将根据典籍引书知识的构造和关系特点，比较知识表示学习和语义网方法，对典籍引书相关的著录知识、引用内容和引用关系进行知识表示和知识建模；将利用以预训练模型为代表的深度学习方法，在典籍语料库的基础上，实现包括实体识别、实体链接和关系抽取的典籍引书知识抽取；将结合已获取的典籍引书知识，比较RDF系统和图数据库两种技术，实现典籍引书知识存储，从而完成典籍引书知识图谱构建。最后，根据构建的知识图谱，项目将结合典籍引书传承影响分析，展示其应用价值和前景。

Ancient Chinese classics carry excellent Chinese traditional culture. Citations from classics are an important guide for discovering the context of traditional culture. Knowledge graph construction represents the technological frontier of the era of artificial intelligence. The project is titled "Research on Construction and Application of Knowledge Graph of Classics' Citations Based on Deep Learning". The representative ancient Chinese digital classics are used as data sources. The classics are constructed on the basis of text preprocessing and manual collation and shallow syntactic analysis. The research will compare the knowledge representation learning and semantic web methods based on the structure and relationship of classics citation knowledge, and perform knowledge representation and knowledge modeling on bibliographical knowledge, reference content, and citation relationships related to classics citations. A deep learning method represented by a pre-trained model will be used to implement knowledge extraction of classics citation books, including entity recognition, entity linking, and relationship extraction, based on the classics corpus. Combining the acquired knowledge of bibliographic references, the two technologies of RDF system and graph database will be compared to achieve the storage of classic bibliographic knowledge, thereby completing the construction of the knowledge graph. Finally, according to the constructed knowledge graph, the project will take the influence analysis of classics as an example to show its application value and prospect.