基于中成药知识图谱的知识推理及智能推荐马宸睿孟子琪边新宇李癑函赵汉青(河北大学中医学院保定071000)〔摘要〕目的/意义构建中成药说明书知识图谱,实现基于知识推理的智能推荐。方法/过程基于Py2neo结合Neo4j图数据库搭建知识图谱的技术,整理1380种中成药信息并构建数据库,基于三元闭包算法实现知识推理,运用概率模型计算中成药推荐评分。结果/结论共形成实体概念11103个,语义关系24种。构建了中成药智能推荐知识图谱,搭建中成药智能推荐平台实现中成药的准确推荐。实现中成药与知识图谱领域结合,为中医辅助诊疗提供方法借鉴,为进一步开展中成药知识可视化研究提供参考。〔关键词〕中成药;知识图谱;数据库;知识推理;智能推荐〔中图分类号〕R-058〔文献标识码〕A〔DOI〕103969/jissn1673-6036202404003KnowledgeReasoningandIntelligentRecommendationBasedonKnowledgeGraphofChinesePatentMedicineMAChenrui,MENGZiqi,BIANXinyu,LIYuehan,ZHAOHanqingCollegeofTraditionalChineseMedicine,HebeiUniversity,Baoding071000,China〔Abstract〕Purpose/SignificanceToconstructtheknowledgegraphofChinesepatentmedicineinstructions,andtorealizeintelligentrecommendationbasedonknowledgereasoningMethod/ProcessBasedonthetechnologyofPy2neocombinedwithNeo4jgraphdatabasetobuildknowledgegraph,theinformationof1380kindsofChinesepatentmedicinearesortedoutandthedatabaseisbuiltKnowledgereasoningisrealizedbasedonthetriadicclosurealgorithm,andtherecommenda...