单篇论文被引频次影响因素及预测研究综述  

A Review of Research on Influencing Factors and Prediction of Citation Frequency of a Single Paper

作　　者：张素芳[1] 刘慧敏 Zhang Sufang;Liu Huimin(School of Economics and Management,South China Normal University,Guangzhou 511400)

机构地区：[1]华南师范大学经济与管理学院,广州511400

出　　处：《知识管理论坛》2022年第3期299-313,共15页Knowledge Management Forum

摘　　要：[目的/意义]梳理单篇论文被引频次的相关影响因素以及被引频次预测研究现状,为科研人员和科研机构研究单篇论文被引频次影响因素及预测提供一个全面系统的认知框架。[过程/方法]采用文献调研法,通过对现有文献进行系统的梳理,总结被引频次预测的影响因素、研究对象和研究方法的相关内容和特点,并通过列表的方式对比分析不同的方法,总结现有研究普遍存在的问题和一些创新的解决方案。[结果/结论]在系统梳理和总结的过程中发现,影响因素与预测结果之间因果关系不明确,研究样本数据缺乏多样性,未明确研究结果的适用性与预测周期的关系,模型评估可解释性较弱。因此,应从解决问题的前提条件、选择有针对性的样本、改进影响因素提取方法、运用数学思维方式进行建模等方面提高后续研究的质量。[Purpose/Significance]Combing the relevant influencing factors of the citation frequency of a single paper and the research status of the prediction of the citation frequency,this paper provides a comprehensive and systematic cognitive framework from the perspective of the involvement of scientific researchers and scientific research institutions in such research.[Method/Process]Using the literature research method,through the systematic combing of the existing literature,this paper summarized the relevant contents and characteristics of the influencing factors,research objects and research methods of citation prediction,compared and analyzed different methods by means of list,and summarized the common problems and some innovative solutions of the existing research.[Result/Conclusion]In the process of systematic combing and summarizing,it is found that the causal relationship between influencing factors and prediction results is not clear,the research sample data is lack of diversity,the relationship between the applicability of research results and prediction cycle is not clear,and the interpretability of model evaluation is weak.Therefore,we should improve the follow-up research quality from the aspects of solving the preconditions of the problem,selecting targeted samples,improving the extraction methods of influencing factors,and using mathematical thinking mode for modeling.

关 键 词：被引频次预测 影响因素 回归分析 机器学习 深度学习 

分 类 号：G353.1[文化科学—情报学]