摘要
[目的/意义]挖掘和组织先秦典籍中的植物知识,构建先秦典籍植物知识图谱,对认识我国古代人民社会和生活状态等具有重要意义。[方法/过程]对先秦典籍中植物词进行详尽标注与计量分析;基于条件随机场(CRF)和多种深度学习模型构建古汉语植物命名实体识别模型,比较分析各模型性能以确定最优模型;设计面向知识图谱的古汉语植物知识组织模式。[结果/结论]基于古汉语预训练语言模型SikuRoBERTa构建的古汉语植物命名实体识别模型性能最优,调和平均值达85.44%,为基于实体的植物知识挖掘提供了有效方法;所构建的先秦典籍植物知识图谱可实现对先秦典籍中植物实体及其关联知识的聚合与可视化呈现。
[Purpose/Significance]The knowledge mining of plants in pre-Qin classics and the construction of pre-Qin plant knowledge graph are of great significance for understanding the society and living conditions of ancient Chinese people.[Method/Process]This paper made a detailed labeling and quantitative analysis of plant words in pre-Qin classics.Based on CRF and a variety of deep learning models,an ancient Chinese plant named entity recognition model was constructed,and the performance of each model was compared and analyzed to de-termine the optimal model.A knowledge graph-oriented knowledge organization model of plants from classics was designed.[Result/Conclusion]The named entity recognition model for ancient Chinese plant based on the ancient Chinese pre-trained language model SikuRoBERTa achieved the best performance,and the harmonic average reached 85.44%,which provided an effective method for entity-based plant knowledge mining.The constructed knowledge graph for pre-Qin classics'plant knowledge can aggregate and visually present plant entities and their relatedknowledge in thepre-Qin classics.
作者
吴梦成
林立涛
齐月
黄水清
王东波
刘浏
Wu Mengcheng;Lin Litao;Qi Yue;Huang Shuiqing;Wang Dongbo;Liu Liu(College of Information Management,Nanjing Agricultural University,Nanjing 210095;Research Center for Humanities and Social Computing,Nanjing Agricultural University,Nanjing 210095;Research Center for Correlation of Domain Knowledge,Nanjing Agricultural University,Nanjing 210095)
出处
《图书情报工作》
北大核心
2023年第12期103-113,共11页
Library and Information Service
基金
国家社会科学基金重大项目“中国古代典籍跨语言知识库构建及应用研究”(项目编号:21&ZD331)
国家自然科学基金青年项目“基于深度学习的典籍引书知识图谱构建及应用研究”(项目编号:72004095)研究成果之一。
关键词
数字人文
先秦典籍
植物命名实体
深度学习
知识图谱
digital humanities
pre-Qin classics
plant named entity
deep learning
knowledge graph