Documentation, Informaiton & Knowledge ›› 2023, Vol. 40 ›› Issue (6): 89-97,116.doi: 10.13366/j.dik.2023.06.089

• Intelligence, Information & Sharing • Previous Articles     Next Articles

Current Advances of Time Series Analysis in Information Science: Tasks, Processes and Problems

CHEN Guo, WANG Kaiyue   

  1. School of Economics and Management, Nanjing University of Technology, Nanjing, 210094
  • Online:2023-11-10 Published:2023-12-10
  • Contact: Correspondence should be addressed to CHEN Guo,Email:delphi1987@qq.com, ORCID:0000-0003-2873-1051
  • Supported by:
    This is an outcome of the project "Research on the Construction of a Disciplinary Research Frontier Identification System Based on Domain Entities"(21YJC870003)supported by Youth Foundation for Humanity and Social Scienceresearch of Ministry of Education of China, and the Youth Project "Research on Domain Knowledge Analysis Models for Frontier Technology Monitoring"(21TQC002)supported by Social Science Foundation of Jiangsu Province.

Abstract: [Purpose/Significance] This paper aims to examine the current research progress in time series analysis within the field of information science and provide a comprehensive overview of common challenges, ultimately serving as a valuable reference for the advancement of modeling and prediction in information science research. [Design/Methodology] This study summarizes the application tasks, research processes and prevalent issues pertaining to time series analysis in information science research, specifically focusing on tasks, processes and problems. [Findings/Conclusion] The findings reveal that, in terms of tasks, existing research has successfully applied time series analysis in various scenarios such as subject theme evolution, academic influence evaluation, network sentiment analysis, and technology trend analysis. The application scope predominantly encompasses historical evolution and future prediction. Regarding the research processes, existing studies primarily follow a sequential order encompassing time series data selection, time segmentation methods, morphological pattern mining, prediction and evaluation. In terms of problems, future research should prioritize the development of time series models for short sequences and enhance the evaluation of time series analysis outcomes. [Originality/Value] Through a comprehensive review, this paper provides a comprehensive overview and reference for researchers in the field by systematically summarizing current research on time series analysis in information science through a comprehensive compendium.

Key words: Time series analysis, Information science research, Subject theme evolution, Academic influence evaluation, Network public opinion analysis, Technology trend analysis