基于主题模型的科技文献主题演化及优化方法研究综述  

A Review of Research on the Improvement of Topic Model Based Topic Evolution Analysis Methods for Scientific Literature

作　　者：于诗睿 李爱花 林紫洛 陈逸菲 唐小利[1] YU Shirui;LI Aihua;LIN Ziluo;CHEN Yifei;TANG Xiaoli(Library&Institute of Medical Information,Chinese Academy of Medical Sciences&Peking Union Medical College,Beijing 100005,China)

机构地区：[1]中国医学科学院/北京协和医学院医学信息研究所/图书馆,北京100005

出　　处：《医学信息学杂志》2023年第8期31-36,共6页Journal of Medical Informatics

基　　金：中国医学科学院医学与健康科技创新工程重大协同创新项目(项目编号:2021-I2M-1-033)。

摘　　要：目的/意义梳理主题演化分析方法研究进展,改善科技文献主题识别和趋势研判效果,支撑文献信息服务,为未来深入研究和实践应用指明方向。方法/过程根据隐含狄利克雷分布主题模型的缺陷归纳其衍生模型,针对现有科技文献主题演化分析方法存在的不足对相应改善方案进行总结,最后结合现有研究局限性提出展望。结果/结论通过全面调研,为主题演化研究提供启示和借鉴。Purpose/Significance The research progress of topic evolution analysis method is sorted out,so as to improve the effect of topic identification and trend judgment of scientific literature,support literature information service,and point out the direction for future in-depth research and practical application.Method/Process The derivative model of latent Dirichlet allocation(LDA)is summed up according to the defects of the topic model,and the corresponding improvement plan is summarized according to the shortcomings of the analysis methods of the topic evolution of the existing scientific literature.Finally,the prospect is put forward according to the limitations of the existing research.Result/Conclusion Through comprehensive investigation,the paper can provide inspiration and references for the study of topic evolution.

关 键 词：主题模型 主题识别 主题演化 隐含狄利克雷分布主题模型 文本挖掘 

分 类 号：R-058[医药卫生]