中国感染控制杂志2026，Vol.25Issue(4)：566-573,8.DOI:10.12138/j.issn.1671-9638.20263102
基于门诊病历的流行性感冒知识图谱构建与应用研究
Construction and application of influenza knowledge graph based on outpa-tient medical records
摘要
Abstract
Objective To compare the performance of different deep learning algorithm models and large language models in the automatic extraction task of influenza entities,and construct knowledge graph in the domain of influ-enza based on the RoBERTa-BiLSTM-CRF model.Methods Entity extraction and model construction were performed based on 4 421 outpatient medical records of influenza from hospital.Through comparison and analysis of the performance of 6 machine learning algorithms and 2 large language models,the optimal model was screened out for triplet extraction.The correlation of entity-relationship pairs was validated with chi-square test.Based on RoBERTa-BiLSTM-CRF model,and combined with medication database matching,similarity algorithm calculation,and manual review,entity alignment was finally achieved.Influenza knowledge graph was constructed using the Neo4j graph database.Results RoBERTa-BiLSTM-CRF model achieved the best performance in the outpatient medical record entity extraction,with an accuracy of 0.931,recall of 0.934,and F1 score of 0.932.Entity catego-ries of influenza-related examinations,symptoms,diseases,medications,and methods were successfully extracted,and 690 effective relationship pairs were found out.Finally,a visualized knowledge graph was constructed based on the extracted influenza entity-related dataset.Conclusion This study validates the significant advantages of the RoBERTa-BiLSTM-CRF model in entity extraction from outpatient medical records,constructs an influenza know-ledge graph based on the model,and provides methodological reference and data foundation for research in medical entity extraction and related fields.关键词
流行性感冒/门诊病历/知识图谱/实体抽取Key words
influenza/outpatient medical record/knowledge graph/entity extraction分类
医药卫生引用本文复制引用
戴萍萍,陈泽华,聂洋波,彭振,姚志强,许林勇..基于门诊病历的流行性感冒知识图谱构建与应用研究[J].中国感染控制杂志,2026,25(4):566-573,8.基金项目
湖南省重点研发计划项目"突发公共卫生事件早期预测预警关键技术攻关"(2023SK2005) （2023SK2005）
中华医学会医学教育研究课题项目"人工智能交叉背景下生物统计学课程研究"(2023B356) （2023B356）
