图书情报工作 ›› 2019, Vol. 63 ›› Issue (19): 128-135.DOI: 10.13266/j.issn.0252-3116.2019.19.013

• 知识组织 • 上一篇    下一篇

基于ERGM的学科交叉领域知识连接机制实证研究

操玉杰1, 李纲1, 毛进1, 杨冠灿2   

  1. 1. 武汉大学信息资源研究中心 武汉 430072;
    2. 中国人民大学信息资源管理学院 北京 100872
  • 收稿日期:2018-12-06 修回日期:2019-04-21 出版日期:2019-10-05 发布日期:2019-10-05
  • 通讯作者: 毛进(ORCID:0000-0001-9572-6709),副研究员,博士后,通讯作者,E-mail:danveno@163.com
  • 作者简介:操玉杰(ORCID:0000-0002-8899-9626),博士研究生;李纲(ORCID:0000-0001-5573-6400),教授,博士生导师;杨冠灿(ORCID:0000-0002-1706-1884),讲师,博士。
  • 基金资助:
    本文系国家自然科学基金青年项目"基于学术异质网络表示学习的知识群落发现"(项目编号:71804135)和中国博士后科学基金项目"融合语义与关系的科研社群识别与演化研究"(项目编号:2018M630885)研究成果之一。

An Empirical Study on Knowledge Connection Mechanism of Interdisciplinary Field Based on ERGM

Cao Yujie1, Li Gang1, Mao Jin1, Yang Guancan2   

  1. 1. Center for Studies of Information Resources, Wuhan University, Wuhan 430072;
    2. School of Information Resource Management, Renmin University of China, Beijing 100872
  • Received:2018-12-06 Revised:2019-04-21 Online:2019-10-05 Published:2019-10-05
  • Supported by:
     

摘要: [目的/意义] 旨在通过探讨学科交叉领域共词网络生成的影响因素及其作用机理,揭示学科交叉领域的微观知识连接机制。[方法/过程] 结合网络嵌入性理论,将学科交叉领域关键词共现关系建立的影响因素归纳为网络结构因素(内生变量)和关键词属性因素(外生变量),进而借助指数随机图模型,选择学科交叉领域"医学信息学"开展实证研究。[结果/结论] 研究结果表明:网络结构对共现关系生成的影响大于关键词本身属性的影响;择优连接机制和传递性机制具有显著正向作用;关键词节点倾向于与较新节点相连;医学信息学的关键词倾向于与基础学科的关键词建立共现关系,而基础学科的关键词却倾向于与自身学科关键词相连。

 

关键词: 指数随机图模型, 跨学科, 同质性, 择优连接, 时间效应

Abstract: [Purpose/significance] The article aims to explore the factors and their mechanisms influencing the generation of co-word network for interdisciplinary field, and to reveal micro-level mechanisms of knowledge connection in interdisciplinary field.[Method/process] Borrowing network embedding theory, the article summarizes the factors into network structure factors (endogenous variables) and keywords' attribute factors (exogenous variables). Exponential random graph model is constructed based on these factors to perform an empirical analysis on the field of Medical Informatics.[Result/conclusion] The results show that the influence of network structure factors on the co-occurrence relationship generation is greater than that of keywords' attributes. Preferential attachment and transitive mechanism have significant positive effect. Keywords tend to be connected with the newer ones. In addition, the keywords of Medical Informatics tend to establish co-occurrence relations with the keywords from basic disciplines, while the keywords from basic disciplines tend to be connected with the keywords in their own disciplines. The conclusions are helpful to understand the formation process of knowledge systems in interdisciplinary fields and the interactions of interdisciplinary knowledge.

Key words: ERGM, interdisciplinary, homophily, preferential attachment, time effect

中图分类号: