图书情报知识 ›› 2022, Vol. 39 ›› Issue (6): 143-157.doi: 10.13366/j.dik.2022.06.143

• 情报、信息与共享 • 上一篇    下一篇

图情领域LDA主题模型应用研究进展述评

张东鑫,张敏   

  • 出版日期:2022-11-10 发布日期:2023-02-24

A Review on Application Studies of LDA Topic Models in Library and Information Science Field

  • Online:2022-11-10 Published:2023-02-24

摘要: [目的/意义]系统归纳梳理LDA模型的应用过程与应用领域,为图情领域LDA模型研究提供参考。[研究设计/方法]以Web of Science核心集、LISA、Google Scholar、中国知网、维普和万方等为数据源,检索图情领域LDA模型的研究文献,通过内容分析构建了LDA模型应用研究分析框架,从模型应用过程的视角对国内外研究现状进行系统地总结归纳。[结论/发现]研究发现,LDA主题模型研究已经形成较为成熟的分析流程,已应用在主题探索、知识组织、学术评价、情感分析等很多领域,但是在应对大数据、多模态数据等复杂处理任务,提升建模结果的语义质量,扩展模型应用等方面还亟待加强。[创新/价值]基于LDA模型的应用过程,细致揭示了图情领域LDA模型应用研究存在的问题和发展方向。

关键词: LDA, 主题建模, 文本挖掘, 图书情报领域

Abstract: [Purpose/Significance] According to previous work, this paper aims to summarize and sorte out the applied process and applied fields of LDA model to provide reference for the research of LDA model in library and information science(LIS)field. [Design/Methodology] We selected Web of Science Core Collection、LISA、Google Scholar、CNKI、VIP and WANFANG Database as data source, retrieved literature about the LDA model in the field of LIS, constructed the analytical framework of LDA model application research through content analysis, from the perspective of the applied process of these models, carefully analyzed the current research at home and abroad [Findings/Conclusion] The results show that a more mature analysis process has been formed for the research of LDA topic models in LIS field, it has been applied in rich fields such as topic exploration, knowledge organization, academic evaluation, sentiment analysis,but the research still need to be strengthened in the future in dealing with complex tasks such as processing big data and multimodal data,improving the semantic quality of modeling results and the application of extended model [Originality/Value] Based on the applied process of LDA model, the existing problems and development direction of LDA model research in the field of LIS are revealed in detail.


Keywords: LDA, Topic modeling, Text mining, LIS field