摘要
【目的】我们使用深度学习模型对于文章进行多分类,研究论文发表机构的学科融合的科研现状。【方法】我们设计了"多类别分类"模型,并应用卷积神经网络对中国科学院产生的8个不同主题的研究论文摘要进行分类。【结果】结果表明,科学研究涉及的学科交叉融合变得日趋紧密。【结论】多学科的融合交叉促进了科研产出,该研究可进一步用于科研机构的战略规划部署和评价等问题。
[Objective]We use deep learning models to multi-classify articles and analyze the disciplinary integration situation of the corresponding institutions.[Methods]In this paper,we design a one-versus-rest classification model and applied convolutional neural networks to categorize paper abstracts of 8 different main subjects produced by Chinese Academy of the Sciences.[Results]The results show that the cross-integration of disciplines involved in scientific research becomes a more frequent practice and the integration of academic fields are promoting the number of publications of scientific research papers.[Conclusions]This research can benefit the strategic planning and deployment for scientific research institutions.
作者
刘晓东
倪浩然
Liu Xiaodong;Ni Haoran(Center of Informatization Strategy and Evaluation,Computer Network Information Center,Chinese Academy of Sciences,Beijing 100190,China;Mathematics for Real-World Systems CDT,University of Warwick,Coventry,United Kingdom)
出处
《数据与计算发展前沿》
2020年第5期99-109,共11页
Frontiers of Data & Computing
基金
supported by the National Science Library of the Chinese Academy of Sciences。
关键词
文本分类
自然语言处理
卷积神经网络
分类算法
text classification
natural language processing
convolutional neural network
classification algorithm
