中医汤剂知识图谱的构建与样例查询方法研究  

Research on Construction of Knowledge Graph of Traditional Chinese Medicine Decoction and Example Query Methodology

作　　者：李思思 燕燕[1] 夏书剑[1] 孙艳秋[1] 丁琳琳[2] LI Sisi;YAN Yan;XIA Shujian;SUN Yanqiu;DING Linlin(Liaoning University of Traditional Chinese Medicine,Shenyang 110847,Liaoning,China;Liaoning University,Shenyang 110036,Liaoning,China)

机构地区：[1]辽宁中医药大学,辽宁沈阳110847 [2]辽宁大学,辽宁沈阳110036

出　　处：《中华中医药学刊》2024年第8期31-36,I0006,I0007,共8页Chinese Archives of Traditional Chinese Medicine

基　　金：国家自然科学基金项目(62072220);辽宁省应用基础研究计划项目(2023JH2/101300054);辽宁省自然科学基金项目(2022-KF-13-06)。

摘　　要：利用现代化信息技术为中医药的科研、传播、教学等方面提供有效的研究方法与途径是当前的研究热点,更是未来的发展方向。知识图谱通过其独特的知识展示,为挖掘医学知识中隐藏的关联关系提供方法,为中医药的传承发展提供动力,为中医药的学习研究提供参考。中医方剂作为中医治疗的重要手段之一,扮演着重要的角色。利用知识图谱技术,实现对中药方剂中汤剂类型的经方进行实体和关系的提取,并构建一个中医汤剂知识图谱。通过对中医文献、中医医案、中医方剂数据库、名老中医经验等非结构化及半结构化信息进行深入分析和文本挖掘,将整理好的知识存入Neo4j图数据库中,并使用中医药物图片作为相应中药实体节点显示,帮助理解知识图谱中的实体和关系并发现隐藏的模式、趋势和关联性。设计一种样例查询算法进行查询分析,为中医汤剂知识的学习、浏览及检索提供新思路,为分析挖掘中医药中潜在的关联关系提供新方案,为中医药知识的表达及可视化提供新范式。The use of modern information technology to provide effective research methods and approaches for scientific research,dissemination and teaching of traditional Chinese medicine is a current research hotspot,but also a future development direction.Knowledge graph,as an effective and novel knowledge management technology,provides a method for mining the hidden correlations in medical knowledge through its unique knowledge representation,gives impetus to the development of the heritage of traditional Chinese medicine,and supplies references for the study and research of traditional Chinese medicine.As an important part of traditional Chinese medicine,the formulae have accumulated a wealth of valuable experience and knowledge,and play an important role as one of the important means of traditional Chinese medicine treatment.The aim of this paper is to realize the extraction of entities and relationships for the decoction types of traditional Chinese medicine formulae and to construct a knowledge graph.Through in-depth analysis and text mining of unstructured and semi-structured information such as Chinese medicine literature,Chinese medicine cases,Chinese medicine prescription databases and experiences of famous and old Chinese medicine practitioners,the organized knowledge is deposited into the Neo4j graph database,and images of Chinese medicines are used to be displayed as nodes of the corresponding medicine entities,which helps users to understand the entities and relationships in the knowledge graph and to discover the hidden patterns,trends and correlations.A example query algorithm was designed for query analysis to provide new ideas for learning,browsing and retrieving knowledge of traditional Chinese medicine formulae,new solutions for analyzing and mining potential associative relationships in traditional Chinese medicine,and new paradigms for expressing and visualizing traditional Chinese medicine knowledge.

关 键 词：知识图谱 中医汤剂 Neo4j 可视化 样例查询 

分 类 号：R289[医药卫生—方剂学]