[目的/意义]从数字人文视角下构建古汉语典籍领域的知识图谱自动问答系统,促进中华传统文化发展与创新。[方法/过程]文章以《左传》为具体研究对象,在此基础上使用支持向量机(SVM)算法实现问句的意图识别,基于BERT-LSTM-CRF的深度学习算法实现问句的实体识别功能,再通过构造Cpyher查询表达式在Neo4j数据库中检索并返回结果;前端则基于Flask框架搭建展示平台供用户使用,最终实现问答系统的搭建。[结果/结论]该问答系统可以实现古文领域问题的智能检索,具有应用价值。
[Purpose/significance]Construct an automatic question-and-answer system of knowledge graphs in the field of ancient Chinese classics from the perspective of digital humanities to promote the development and innovation of traditional Chinese culture.[Method/process]Taking "Zuo Zhuan" as the specific research object,on this basis,support vector machine (SVM) algorithm is used to realize the intention recognition of question sentences,and the deep learning algorithm based on BERT-LSTM-CRF realizes the entity recognition function of question sentences.Then cpyher query expression is constructed to retrieve and return the results in Neo4j database; the front page builds a display platform based on Flask framework for users to use,and finally realizes the construction of question and answer system.[Result/conclusion]The question answering system can realize intelligent retrieval of questions in the field of ancient Chinese and has application value.