基于潜在狄利克雷分布主题模型的初产妇产后健康信息需求研究  

Postpartum health information need of nulliparous women based on latent Dirichlet allocation

作　　者：郭赛男 蒋慧萍 王子豪 梁秋曼 史婷奇[1] GUO Sai-nan;JIANG Hui-ping;WANG Zi-hao;LIANG Qiu-man;SHI Ting-qi(Dept.of Nursing Administration,Nanjing Drum Tower Hospital,the Affiliated Hospital of Nanjing University Medical School,Nanjing 210008,China;School of Nursing,Nanjing University of Chinese Medicine,Nanjing 210008,China)

机构地区：[1]南京大学医学院附属鼓楼医院护理部,江苏南京210008 [2]南京中医药大学护理学院,江苏南京210008

出　　处：《护理学报》2024年第19期19-23,共5页Journal of Nursing（China）

基　　金：南京市卫生科技发展专项资金项目(YKK22074);南京大学中国医院改革发展研究院课题项目(NDYG2022067)。

摘　　要：目的运用潜在狄利克雷分布(Latent Dirichlet Allocation,LDA)主题模型深入挖掘即时社交平台产后母婴保健信息需求。方法2023年1—6月提取产后延续性护理微信群内文本数据,通过数据清洗、分词和LDA主题模型构建,分析文本数据所蕴含的需求主题。结果LDA主题模型将所提取的23531条文本数据划分为8个主题:婴儿健康状况、婴儿喂养状况、婴儿日常护理、生长发育、母婴健康体检、疫苗接种、产后恢复、社会支持和同伴经验分享。结论基于自然语言的信息需求分析能获取客观全面的产后母婴健康信息需求,为医疗机构开展全面、精细化的产后健康指导提供参考。Objective To explore health care information demand for postpartum women and infants on real-time social platforms using the latent Dirichlet allocation(LDA).Methods The text data in the WeChat group for postpartum continuous nursing were extracted from January to June 2023,and the demand topics contained in the text data were analyzed through data cleaning,word segmentation and LDA.Results The 23531 text data extracted by LDA were summarized into 8 topics:infant health status,infant feeding status,infant daily care,growth and development,maternal and infant health examination,vaccination,postpartum recovery,social support,and peer experience sharing.Conclusion Information need analysis based on natural language is effective for obtaining objective and comprehensive health information needs for postpartum maternal and infant,and provide reference for medical institutions to provide comprehensive and refined postpartum health guidance.

关 键 词：初产妇 产后 母婴健康信息需求 延续性护理 潜在狄利克雷分布(LDA)主题模型 

分 类 号：R473.71[医药卫生—护理学]