【作者】 刘琳；

【导师】 宋胜利； 南敏庚；

【作者基本信息】 西安电子科技大学 ， 工程硕士（专业学位）， 2021， 硕士

【摘要】 随着人们生活水平的提高,高质量的互联网医疗服务成为目前社会的一项迫切需求。“互联网+医疗”的普及,催生了一些医疗服务平台,但是多数平台都是以互联网公司为主导构建的服务平台,缺乏专业医院的参与,并且用药信息查询、在线咨询、项目检测预约、在线查看检测结果等患者的实际需求得不到有效解决。目前也存在一些社区网站能够提供在线咨询、就诊功能,但是大多在线服务平台由于不够自动化和智能化,无法为大量患者提供及时准确的个性化用药服务。针对上述问题,本文在对已有的医药服务平台调研的基础上,结合实际项目应用需求,以某医院药剂科提供的医药数据为数据源,研究并实现了基于知识图谱的精准用药智能服务平台。所研发平台采用Spring框架作为后端逻辑开发框架,以微信小程序作为客户端交互平台,采用关系型数据库My SQL存储患者基本信息和就医信息,利用Neo4J来存储本文的医药知识库,构建了精准用药服务平台的知识图谱。基于知识图谱的精准用药服务平台主要划分为四个主要模块:症状上报模块、药物浓度检测模块、在线咨询模块和数据分析模块:(1)利用模糊匹配方法实现了症状上报模块的药物名称规范,利用Jieba分词、设置停用词表和分词词典、短语拼接等技术完成文本关键词和短语的抽取工作,实现医药领域知识图谱更新、患者症状上报和查看;(2)在药物检测模块,利用命名实体识别技术,实现了患者药物名称的识别,在数据查询该药物信息完成药物信息的推送、用户药物浓度检测及结果查看;(3)利用基于词典的正向最大匹配算法完成患者咨询模块的问句分词、关键词提取,利用余弦相似度计算方法实现问句分类,利用LSTM-CRF模型实现问句医疗实体识别,利用哈工大LTP依存句法分析器实现问句的依存句法分析,最终获取问句三元组;(4)利用Cypher查询语言制定知识图谱查询模板,结合问句三元组实现知识图谱查询;(5)利用Echarts工具,实现数据的可视化分析。完成平台的设计实现工作后,该平台根据功能性需求和非功能性需求完成了平台测试,精准用药智能服务平台的各项功能能够正常工作,达到预期结果。通过收集患者提供的用药信息来更新知识图谱,为药剂科的药理研究提供数据支持;基于医药知识图谱,为患者提供高效精准的在线咨询。该平台为患者和药剂师提供的症状上报、查看历史上报记录、预约药物浓度检测、查看历史检测结果等功能能够满足日常需求。更多还原

【Abstract】 With the improvement of people’s living standards,high-quality Internet medical services have become an urgent need of the current society.The popularity of "Internet + medical" has given birth to some medical service platforms,but most of them are service platforms built by Internet companies,lacking the participation of professional hospitals,and drug information inquiry,online consultation,project testing appointments,and online viewing The actual needs of patients,such as test results,cannot be effectively addressed.At present,there are also some community websites that can provide online consultation and medical treatment functions,but most of the online service platforms cannot provide timely and accurate personalized medication services for a large number of patients due to insufficient automation and intelligence.In response to the above problems,based on the investigation of existing medical service platforms,this article combines actual project application requirements and uses medical data provided by a hospital’s pharmacy department as the data source to study and implement a knowledge map-based intelligent drug service platform for precise medication..The R&D platform adopts the Spring framework as the back-end logic development framework,the We Chat applet as the client interaction platform,the relational database My SQL to store basic patient information and medical information,and Neo4 J to store the medical knowledge base of this article to build precise medication The knowledge graph of the service platform.The precision medication service platform based on the knowledge graph is divided into four main modules: symptom reporting module,drug concentration detection module,online consultation module and data analysis module:(1)Fuzzy matching method is used to implement the drug name specification of the symptom reporting module.Use Jieba word segmentation,setting stop vocabulary and word segmentation dictionary,phrase splicing and other technologies to complete the extraction of text keywords and phrases,realize the knowledge map update of the medical field,report and view patient symptoms;(2)In the drug detection module,use naming The entity recognition technology realizes the identification of the patient’s drug name.The drug information is queried in the data to complete the push of drug information,the user’s drug concentration detection and the result view;(3)Use the dictionary-based forward maximum matching algorithm to complete the question of the patient consultation module Sentence segmentation and keyword extraction,use cosine similarity calculation method to realize question classification,use LSTM-CRF model to realize question medical entity recognition,use Harbin Institute of Technology LTP dependency syntax analyzer to realize question dependency analysis,and finally obtain question three Tuples;(4)Use Cypher query language to develop knowledge graph query templates,and combine question triples to realize knowledge graph query;(5)Use Echarts tools to realize data visualization analysis.After completing the design and implementation of the platform,the platform completed the platform test according to the functional requirements and non-functional requirements,and the various functions of the precision medicine intelligent service platform can work normally and achieve the expected results.Update the knowledge map by collecting medication information provided by patients,and provide data support for pharmacological research in the pharmacy department;based on the medical knowledge map,provide patients with efficient and accurate online consultation.The platform provides patients and pharmacists with functions such as reporting symptoms,viewing historical reporting records,making appointments for drug concentration testing,and viewing historical test results to meet daily needs.更多还原