Detecting research topic trends by author-defined keyword frequency
作者：
Highlights：
• This paper proposed author-defined keyword frequency prediction (AKFP) for detecting research topic trends and also offered the practical guidelines and other potential applications of the AKFP such as emerging topics identification.
• The feasibility of short-, medium-, and long-term word frequency prediction was verified systematically, and AKFP achieved excellent performance in computer science field.
• Four categories of proposed features (i.e. Temporal feature, Persistence, community size, and community development potential) have a significant but inconsistent impact on future keyword frequency.
• We proposed a simple but effective method to build a balanced and sufficient training set for the AKFP, when encountering uneven data distribution (e.g. power-law).
摘要
•This paper proposed author-defined keyword frequency prediction (AKFP) for detecting research topic trends and also offered the practical guidelines and other potential applications of the AKFP such as emerging topics identification.•The feasibility of short-, medium-, and long-term word frequency prediction was verified systematically, and AKFP achieved excellent performance in computer science field.•Four categories of proposed features (i.e. Temporal feature, Persistence, community size, and community development potential) have a significant but inconsistent impact on future keyword frequency.•We proposed a simple but effective method to build a balanced and sufficient training set for the AKFP, when encountering uneven data distribution (e.g. power-law).
论文关键词：Scientometrics,Bibliometrics,Deep learning,Word frequency prediction
论文评审过程：Received 31 December 2020, Revised 10 March 2021, Accepted 12 March 2021, Available online 26 March 2021, Version of Record 26 March 2021.
论文官网地址：https://doi.org/10.1016/j.ipm.2021.102594
