Abstract: Background　Biclustering is a more and more widely used data mining technique in recent years,and provides a more scientific analysis method for document management.Objective　To analyze the research hotspots of general medicine,deepen digging the research themes and find the problems in general medical research by using biclustering method.Methods　On March 1st,2016,journal literatures on general medicine were retrieved based on the data resources of China Academic Journals Full-text Database.Bibliographic Item Co-Occurrence Matrix Builder （BICOMB） 2.0 was used to filter prolific authors and high-frequency keywords,and gCLUTO 1.0 was used to conduct biclustering analysis.Results　A total of 2 119 references were retrieved,4 075 authors and 9 155 keywords were obtained.Eventually 64 prolific authors and 59 high-frequency keywords were obtained.According to the actual clustering effects,the authors were clustered into five categories:cluster 0 included authors Jun Shao,Jing-quan Li and Kidd Michael,and the research topic was the development of the global family medicine;cluster 1 included 12 authors such as Feng Xiao,Wei-liang Lin,Li-jun Du,etc.and research topics included the cultivation of family doctors,family doctor system research and the experience exploration of family doctor service;cluster 2 included 33 authors such as Kan Zhang,Rui-jie Gong,Xue-ping Du,etc.and research topic was the cultivation of general medical talents,which was mainly divided into two aspects:general medicine education and on-the-job training of general practitioners;cluster 3 included 8 authors such as Man-xia Li,Wei-hong Ma,Qin Li,etc.and the research topic was on-the-job training and standardization training of general practitioners;cluster 4 included 8 authors such as Wan-xia Huang,Xin-wang Wang,Li-wei Zhang,etc.and research topics included the signing service of family doctors,and general practitioners’ teacher training.Conclusion　The current domestic researches fields of general medicine can be divided into five hot research topics:the development of global family medicine,general medical education,training of general practitioners,family doctor service and system research of family doctors;and the research themes are increasingly detailed and deepened,but there are still many problems in the goal of academic education of general medicine,qualified teachers in general medicine,the basic and deep level research on family doctor systems.

Key words: General practice, Cluster analysis, Hotspots, Author, Keywords

摘要： 背景　双向聚类是近年来应用越来越广泛的一种数据挖掘技术，为文献管理提供了一种较为科学的分析方法。目的　采用双向聚类法分析全科医学研究热点，深入挖掘研究主题，发现存在的问题。方法　2016-03-01，以中国学术期刊网络出版总库为数据来源，检索以全科医学为主题的期刊文献。利用书目信息共现挖掘系统（BICOMB）2.0选取高产作者和高频关键词，采用gCLUTO 1.0软件进行双向聚类分析。结果　共检索到2 119条文献。共得作者4 075位，关键词9 155个。最终取高产作者64位，高频关键词59个。根据实际聚类效果最终选择将作者聚为5类：聚类0作者包括邵钧、李井泉和Michael Kidd；研究主题为全球家庭医学的发展。聚类1作者包括肖峰、林伟良、杜丽君等12位；研究主题包含对家庭医生的培养、家庭医生制度的研究和家庭医生式服务的经验探索。聚类2作者包括张勘、龚睿婕、杜雪平等33位；研究主题为全科医学人才的培养，主要分为全科医学教育和全科医生岗位培训2个方面。聚类3作者包括李曼霞、马维红、李勤等8位；研究主题为全科医师岗位培训和规范化培训。聚类4作者包括黄婉霞、王心旺、张立威等8位；研究主题包括家庭医生签约服务、全科医师师资培训等。结论　目前国内全科医学研究领域可以划分为5个热点研究主题，分别为全球家庭医学的发展、全科医学教育、全科医生培训、家庭医生式服务和家庭医生制度研究，且研究主题越来越细化和深入，但在全科医学学历教育的目标、合格的全科医学师资、家庭医生制度的基础性研究和深层次研究上仍存在着诸多问题。

关键词: 全科医学, 聚类分析, 热点, 作者, 关键词