基于文本挖掘的北京市属公立医院负面舆情研究  

Study on negative public opinion of Beijing public hospitals based on text mining

作　　者：杜孟凯 王蕾[1] 邸洋 单玥 岳小林 DU Mengkai;WANG Lei;DI Yang;SHAN Yue;YUE Xiaolin(Propaganda Center,Xuanwu Hospital Capital Medical University,Beijing100053,China;Department of Party-Masses,Beijing Hospitals Authority,Beijing100053,China;Party Committee Office,Xuanwu Hospital Capital Medical University,Beijing100053,China)

机构地区：[1]首都医科大学宣武医院宣教中心,北京100053 [2]北京市医院管理中心党群处,北京100053 [3]首都医科大学宣武医院党委办公室,北京100053

出　　处：《中国医药导报》2023年第9期190-193,共4页China Medical Herald

基　　金：北京市医院管理中心“培育计划”项目(PG2022017)。

摘　　要：目的本研究基于文本挖掘与主题建模的方法对北京市属公立医院负面舆情进行分析。方法收集北京市22家市属公立医院2021年1月1日至12月31日在主流网媒平台的全部负面舆情。采用Python 3.9进行文本挖掘并建立语料库,利用隐含狄利克雷分布(LDA)主题模型对语料进行聚类,并根据每个聚类的关键词解释聚类含义。结果共收集语料3083条,提取6个主题,根据关键词形成号源与资费相关问题、突发公共卫生事件相关问题、等待时间与服务可及性问题、服务态度与就诊流程体验问题、患者入院手术体验与预后问题等5类问题。结论公立医院的负面舆情是反映社情民意的重要信息来源。通过对负面舆情的文本挖掘研究,可以有效发现管理漏洞,提出改进措施,从而提升医院舆情管理工作,改善医疗服务工作。Objective To analyze the negative public opinions of municipal public hospitals in Beijing based on text mining and topic modeling.Methods All the negative public opinions of 22 municipal public hospitals in Beijing on the mainstream online media platforms from January 1 to December 31,2021 were collected.Python 3.9 was used for text mining and corpus building.While the latent dirichlet allocation(LDA)topic model was adopted for linguistic data clustering and explaining the meaning of each cluster according to its keywords.Results A total of 3083 pieces of linguistic data were collected,and 6 topics were extracted,forming 5 types of problems by keywords,including problems related to registration and fees,problems related to public health emergencies,problems with waiting times and service accessibility,problems in service attitude and experience of the medical process,and problems in patients’experience of surgery and prognosis as well as hospital admissions.Conclusion Negative public opinions about public hospitals are an important source of information reflecting social conditions and public opinions.Through text mining and research on negative public opinions,management loopholes can be found effectively and improvement measures proposed,so as to enhance public opinion management in hospitals and improve medical services.

关 键 词：公立医院 舆情管理 文本挖掘 聚类分析 

分 类 号：R197.32[医药卫生—卫生事业管理]