摘要
目的:探索海洋中药数据挖掘和知识补全技术,挖掘海洋中药文献的隐藏知识,为古籍文献的中医药知识挖掘和知识图谱质量提升提供借鉴。方法:针对领域知识图谱的特点提出了一种融合知识图谱与数据挖掘的领域知识图谱补全方法。从海洋中药知识图谱中检索补肾类海洋中药及相关效方,利用关联规则、聚类对海洋中药功效、主治、效方配伍、性味归经等之间的关联关系进行分析与聚类,对海洋中药知识图谱进行知识补全。结果:检索出补肾类海洋中药51味,相关效方123首,以此为数据源进行数据挖掘。发现补肾类海洋中药大多味咸,性温,归肝、肾两经。筛选出83首补肾类效方,发现补肾常用海洋中药有海马、海狗肾、海参、石首鱼和4类核心药组。将挖掘结果以三元组的形式补充到海洋中药知识图谱中,可有效完善补肾类效方的核心配伍关系。结论:以知识图谱为数据源进行数据挖掘,并将挖掘结果补充知识图谱,可以使知识图谱更为完整,相关思路与方法可为中医药领域数据挖掘、中医药知识图谱质量提升提供参考。
Objective:To explore the data mining and knowledge completion technology of marine traditional Chinese medicine,and to explore the hidden knowledge of marine traditional Chinese medicine literature,so as to provide reference for the knowledge mining and knowledge graph quality improvement of ancient Chinese medicine literature.Methods:According to the characteristics of domain knowledge graph,a domain knowledge graph completion method combining knowledge graph and data mining was proposed.The kidneytonifying marine traditional Chinese medicine and related effective prescriptions were retrieved from the knowledge graph of marine traditional Chinese medicine.The association rules and clustering were used to analyze and cluster the correlation between the efficacy,indications,compatibility of effective prescriptions,and meridian tropism of marine traditional Chinese medicine,so as to complete the knowledge graph of marine traditional Chinese medicine.Results:51 kinds of marine Chinese medicine for tonifying kidney and 123 related effective prescriptions were retrieved,which were used as data sources for data mining.It was found that most of the kidney-tonifying marine Chinese medicines are salty,sweet,warm,and belong to the two meridians of liver and kidney.A total of 83 kidney-tonifying effective prescriptions were screened out,and it was found that the commonly used marine Chinese medicines for kidney-tonifying were hippocampus,sea dog kidney,sea cucumber,Sciaenops ocellatus and 4 core drug groups.The mining results are supplemented to the knowledge graph of marine traditional Chinese medicine in the form of triples to effectively improve the core compatibility relationship of kidney-tonifying prescriptions.Conclusion:Data mining with knowledge graph as data source and supplementing knowledge graph with mining results can make knowledge graph more complete.Relevant ideas and methods can provide reference for data mining in the field of traditional Chinese medicine and quality improvement of knowledge graph of traditional Chinese medicine.
作者
洪海蓝
王书怡
李文林
杨丽丽
HONG Hailan;WANG Shuyi;LI Wenlin;YANG Lili(Nanjing University of Chinese Medicine,Nanjing 210023,China;Jiangsu Provincial Engineering Research Center of TCM External Medication Development and Application,Nanjing 210023,China)
出处
《中医药信息》
2023年第10期27-34,共8页
Information on Traditional Chinese Medicine
基金
江苏省中医药科技发展计划(ZT202101)
江苏省研究生科研创新计划(KYCX22_2052)。
关键词
知识图谱
知识补全
海洋中药
数据挖掘
Knowledge graph
Knowledge completion
Marine Chinese medicine
Data mining
