关键词: 图书情报学, 学术观点, 学术思想, 发展轨迹, 深度学习, 学术全文本, 主题模型

Abstract: [Purpose/Significance] The aim of this paper is to present the development track of the academic thoughts and viewpoints in the field of Library and Information Science(LIS),which will not only help to sort out the development process of LIS from a diachronic perspective, but also helps to explore the achievements of LIS from a microscopic perspective [Design/Methodology] This study begins with the definition and review of the academic thoughts and academic viewpoints based on the core journal literature in CNKI of LIS from 1992 to 2022It goes on to formulate labeling specifications for academic thoughts and academic viewpoints and complete labeling manually.Then it builds an automatic extraction model for academic thoughts and academic viewpoints based on deep learning. Finally, combined with the topic model,it systematically explore the development track of academic thoughts and academic viewpoints of LIS. [Findings/Conclusion]The representative themes of academic thought of LIS in each period are "Nguyen's classification thought, Bibliography  ideology,citation analysis, knowledge management, etc.", while the representative themes of academic viewpoint are" reader service, cognitive information science and subject librarian, public opinion evolution, etc.". The academic thoughts and viewpoints of each period correspond to the themes of the development of LIS disciplines in this period. It shows that on the one hand, academic thoughts and their viewpoints play a fundamental role in the development of disciplines and are stable; on the other hand, the emergence of new problems drive the introduction of thoughts and viewpoints of other disciplines and the emergence of new thoughts and viewpoints, which are dynamic and changeable. [Originality/Value] Revealing the development of the academic thoughts and viewpoints of LIS based on data-driven method not only makes up for the shortcomings of traditional methods, but also provides new ideas for scholars to understand the historical evolution and future prediction of LIS.

Keywords: Library and Information Science, Academic thought, Academic viewpoint, Development track, Deep learning, Academic full text, Topic model