图书情报工作 ›› 2016, Vol. 60 ›› Issue (15): 76-85,93.DOI: 10.13266/j.issn.0252-3116.2016.15.011

• 情报研究 • 上一篇    下一篇

语义分类的学科主题演化分析方法研究——以我国图书情报领域大数据研究为例

刘自强, 王效岳, 白如江   

  1. 山东理工大学科技信息研究所 淄博 255049
  • 收稿日期:2016-05-23 修回日期:2016-07-16 出版日期:2016-08-05 发布日期:2016-08-05
  • 通讯作者: 王效岳(ORCID:0000-0002-7100-7758)教授,博士,硕士生导师
  • 作者简介:刘自强(ORCID:0000-0003-1814-8655),硕士研究生;白如江(ORCID:0000-0003-3822-8484)副教授,博士
  • 基金资助:

    本文系国家社会科学基金项目“未来新兴科学研究前沿识别研究”(项目编号:16BTQ083)和教育部人文社会科学研究青年基金项目“基于引文内容分析的科技创新路径识别研究”(项目编号:16YJC870008)研究成果之一。

Research on the Discipline Topic Evolution Analysis Method of Semantic Classification——A Case Study of Big Data in the Field of Library and Information Science in China

Liu Ziqiang, Wang Xiaoyue, Bai Rujiang   

  1. Institute of Scientific & Technical Information,Shandong University of Technology, Zibo 255049
  • Received:2016-05-23 Revised:2016-07-16 Online:2016-08-05 Published:2016-08-05

摘要:

[目的/意义] 学科主题演化研究有助于掌握学科发展现状、研究热点、研究前沿和发展趋势等情况,是进行科技创新的基础,是面向科技创新的重要研究方向。[方法/过程] 提出一种语义分类的学科主题演化分析方法:将关键词分为研究问题、研究方法和研究技术3类,构建不同语义分类的共词网络;然后基于Fast Unfolding社区发现算法识别具有语义特征的社区(主题);利用相似度算法计算相邻子时期主题间的相似度,构建学科主题演化图谱,以分析某学科领域研究问题、研究方法和研究技术的变化,实现深度、细致的学科主题演化分析。[结果/结论] 通过对2012-2015年CNKI数据库收录的我国大数据研究领域相关论文数据的处理分析,证明该方法的准确性和有效性。

关键词: 主题演化, 语义分类, 社区发现算法, 可视化

Abstract:

[Purpose/significance] Topic evolution research is helpful to master the development situation, research focus, research front and development trend in a discipline. It is an important research direction for science and technology innovation.[Method/process] This paper put forward an analysis method of discipline topic evolution based on semantic classification. First, the authors classified keywords into three types of semantic roles——research problems, research methods and research techniques, then constructed co-keywords network of different semantic roles. After this, based on the fast unfolding community discovery algorithm, the authors detected the semantic features of the community (topic); the similarity algorithm was used to calculate similarity between the topics of adjacent periods. The authors also constructed the topic evolution map to analyze changes in a discipline's research problems, research methods and research techniques and achieve an in-depth and meticulous topic evolution analysis. The accuracy and effectiveness of the method are proved according to the analysis of the data of big data in the related research field from 2012-2015 in CNKI.

Key words: topic evolution, semantic classification, community discovery algorithm, visualization

中图分类号: