基于Web of Science数据库的2型糖尿病机器学习可视化文献计量分析  

Bibliometric analysis of machine learning in type 2 diabetes mellitus based on Web of Science

作　　者：杨燕菱 林倍思[1] 刘芷谷 陈亚兰 陈丹蕊 许雯[1] Yang Yanling;Lin Beisi;Liu Zhigu;Chen Yalan;Chen Danrui;Xu Wen(Department of Endocrinology and Metabolism,Guangdong Provincial Key Laboratory of Diabetology,Third Affiliated Hospital of Sun Yat-sen University,Guangzhou 510630,China;Department of Endocrinology and Metabolism,Guangzhou First People′s Hospital,Guangzhou 510180,China)

机构地区：[1]中山大学附属第三医院内分泌与代谢病学科、广东省糖尿病防治重点实验室,广州510630 [2]广州市第一人民医院内分泌与代谢病学科,广州510180

出　　处：《中华糖尿病杂志》2025年第3期360-368,共9页CHINESE JOURNAL OF DIABETES MELLITUS

基　　金：广东省自然科学基金(2022A1515012364)。

摘　　要：目的基于Web of Science核心合集(WoSCC)数据库,利用文献计量学探讨机器学习在2型糖尿病(T2DM)研究领域的现状、热点和未来趋势。方法检索2010—2023年WoSCC的科学引文检索扩展版(SCIE)数据库中收录的关于机器学习与T2DM的研究论文。使用Excel、CiteSpace、VOSviewer、R软件包"bibliometrix"和文献计量在线分析平台对发文量、国家、研究机构、作者、关键词及相关引文进行可视化分析。结果共纳入636篇文献。2021—2023年,机器学习在T2DM领域中的研究论文发文量急剧增加。美国是该领域中参与国际合作最多的国家。哈佛大学是研究论文发表数量最多的机构。共被引文献聚类标签显示了7个主要聚类:筛查、临床支持决策系统、糖尿病肾脏病、肠道微生物组、轻度认知障碍、医疗诊断及差异表达基因。关键词爆发分析表明,风险预测、预测模型、代谢组学和干预措施是研究强度较高的热点。结论机器学习在T2DM领域的相关研究活跃度日益增加。当前该领域的研究热点主要集中在T2DM的预测及筛查方面,而机器学习在T2DM管理措施的应用及对并发症预测的研究可能会成为未来的趋势。Objective To discuss the current status,hot topics in research,and future trends in the field of machine learning for type 2 diabetes mellitus(T2DM)from 2010 to 2023 based on Web of Science.Methods Articles related to machine learning for T2DM from 2010 to 2023 were retrieved from the Science Citation Index-Expanded database within the Web of Science Core Collection.The retrieved data,including the number of published articles,countries,research institutions,authors,keywords,and related citations,were analyzed using Excel,CiteSpace,VOSviewer,the R software package"bibliometrix",and the online analysis platform of literature metrology.Results A total of 636 articles were collected.The number of publications has increased dramatically from 2021 to 2023.The United States was the country most frequently involved in international collaborations.Harvard University emerged as the most productive institution.The co-citation keyword clustering analysis revealed 7 main clusters:screening,decision support systems,diabetic kidney disease,gut microbiome,mild cognitive impairment,medical diagnosis,differentially expressed genes.Keyword burst analysis indicated that risk prediction,predictive model,metagenome and interventions were the research hot topics with high strength.Conclusion Machine learning is now being increasingly applied in the fields of T2DM.Currently,the hot topics in this field are mainly focused on the prediction and screening of T2DM,while intervention and the prediction of complications,may become future research trends.

关 键 词：糖尿病 2型 机器学习 文献计量学 可视化分析 

分 类 号：R587.1[医药卫生—内分泌]