图书情报工作 ›› 2018, Vol. 62 ›› Issue (14): 62-71.DOI: 10.13266/j.issn.0252-3116.2018.14.008

• 情报研究 • 上一篇    下一篇

基于主题变迁的领域发展路径智能化识别——以人工智能为例

周源1, 张超2, 唐杰3, 刘宇飞4, 张宇韬3   

  1. 1. 清华大学公共管理学院 北京 100084;
    2. 华中科技大学机械科学与工程学院 武汉 430074;
    3. 清华大学计算机系 北京 100084;
    4. 中国工程院战略咨询中心 北京 100088
  • 收稿日期:2017-11-21 修回日期:2018-03-14 出版日期:2018-07-20 发布日期:2018-07-20
  • 通讯作者: 刘宇飞(ORCID:0000-0001-9420-8811),博士后,通讯作者,E-mail:liuyufei0418@qq.com
  • 作者简介:周源(ORCID:0000-0002-9198-6586),副教授,博士生导师;张超(ORCID: 0000-0001-7612-9327),硕士研究生;唐杰(ORCID: 0000-0003-3487-4593),副教授,博士生导师;张宇韬(ORCID:0000-0002-5759-1230),博士研究生。
  • 基金资助:
    本文系国家自然科学基金"支持技术预见的多源异构大数据融合与时序文本预测方法研究"(项目编号:91646102)和国家自然科学基金"面向 2035 的中国工程科技发展路线图绘制理论与方法研究"(项目编号:L1624045)研究成果之一。

Intelligent Identification of Field Development Trajectory Based on Topic Evolution: A Case Study of Artificial Intelligence

Zhou Yuan1, Zhang Chao2, Tang Jie3, Liu Yufei4, Zhang Yutao3   

  1. 1. School of Public Policy and Management, Tsinghua University, Beijing 100084;
    2. School of Mechanical Science and Engineering, Huazhong University of Science and Technology, Wuhan 430074;
    3. Department of Computer Science and Technology, Tsinghua University, Beijing 100084;
    4. The CAE Center for Strategic Studies, Chinese Academy of Engineering, Beijing 100088
  • Received:2017-11-21 Revised:2018-03-14 Online:2018-07-20 Published:2018-07-20

摘要: [目的/意义]识别领域发展路径对于科技创新具有重要意义,但现有方法如专家访谈、引文分析等不能适应文献爆发性增长的现状,针对这一问题,提出一种基于主题变迁的领域发展路径识别方法。[方法/过程]该方法可以自动从Aminer平台获取数据,通过构建关键词-学者矩阵,综合使用KMeans++和谱聚类算法识别出研究主题和相关学者;通过相似度计算实现不同主题之间的关联,最终获得研究领域的发展路径并进行可视化展示。[结果/结论]通过对人工智能领域的实证分析,结果表明该方法能够有效反映领域研究主题的变迁,有助于研究者快速定位领域的研究热点和重点,丰富领域发展路径相关的研究方法。

关键词: 领域发展路径, 主题变迁, KMeans++, 谱聚类, 人工智能

Abstract: [Purpose/significance] Identifying the trajectory of development is of great importance to scientific and technological innovations. However, existing methods such as expert interviews and citation analysis cannot meet the current situation of the explosive growth of literature. In response to this problem, this paper proposes a new identification method of filed development trajectory. [Method/process] This method identifies the research topics and related scholars by using Kmeans ++ and spectral clustering algorithms with the keyword-scholar matrix, calculates the correlation between different topics, and finally visualizes the trajectory of developmen. [Result/conclusion] Through the empirical analysis of the field of artificial intelligence, the results show that the method can effectively reflect the evolution of the topic of field research, help researchers quickly locate popular research topics and focuses, and enrich the research methods related to the trajectory of field development.

Key words: field development trajectory, topic evolution, KMeans++, spectral clustering, artificial intelligence

中图分类号: