应用共词分析法揭示生物医学工程领域的研究主题

Revealing Theme Structure of Biomedical Engineering Using Co-Word Analysis

基于期刊论文关键词,探讨共词分析法揭示生物医学工程领域研究热点及其主题结构的适用性。首先利用CNKI和万方数据知识服务平台,检索《国际生物医学工程杂志》、《生物医学工程学杂志》和《中国生物医学工程学报》3种期刊自创刊到2011年12月5日刊载的中文文献,得到有效论文11 574篇,共包含34 221个关键词。然后经过数据清洗,从中提取出38个高频关键词,总频次为16 968次。并以此为基础,利用TDA软件制作关键词共词矩阵,将其导入SPSS和Ucinet软件生成聚类和共词网络图谱。最后结合聚类和共词网络图谱将主要研究内容分为5个方面,其中生物材料、人工器官与组织工程领域包含12个高频关键词,占高频关键词总频次(简称总频次)的34.7%;生物力学领域包含7个高频关键词,占总频次的11.4%;生物信号及医学图像领域包含9个高频关键词,占总频次的25.7%;信号测量及超声诊断应用领域包含5个高频关键词,占总频次的12.1%;其它领域包含5个高频关键词,占总频次的16.1%。研究结果可较为客观地反映生物医学工程领域的研究热点及其主题结构,为科研人员分析生物医学工程领域及其子领域的主题结构及研究趋势提供一些思路。

This paper aims to explore the applicability of co-word analysis in revealing research hotspots and theme structure of BME based on journal publications. In the first step, 11 574 articles in Chinese publicated in all issues by December 5th, 2011 in the three Journals including International Journal of Biomedical Engineering, Journal of Biomedical Engineering and Chinese Journal of Biomedical Engineering were collected from CNKI and WANFANG DATA. By data cleaning, 38 high-frequency keywords, with total frequency of 16 968, were extracted among 34 221 keywords in the articles. Based on these 38 high-frequency keywords, coword matrix, clustering dendrogram and co-occurrence network diagram were constructed by data processing and analyzing tools including TDA, SPSS18.0 and Ucinet. After that based on clustering dendrogram and cooccurrence network diagram, five main research themes were obtained as biological materials, artificial organs and tissue engineering, these words contained 12 high-frequency keywords and their total frequency accounted for 34.7% of 16 968 ; biomechanics which contained seven high-frequency keywords and their total frequency accounted for 11.4% of 16 968 ; biomedical signal and medical image which contained nine high-frequency keywords and their total frequency accounted for 25.7% of 16968; application of signal measurement and ultrasound diagnosis which contained five high-frequency keywords and their total frequency accounted for 12. 1% of 16968; others which contained five high-frequency keywords and their total frequency accounted for 16. 1 percent of 16968. The results can properly reflect research hotspots and theme structure of BME, providing some ideas to identify research hotspots and theme structure in BME and its sub-fields.