基于BERTopic算法的引文主题实证分析——以一篇高被引诺贝尔生理学或医学奖论文为例  

Empirical Analysis of Citation Topic Based on BERTopic:A Case Study of a Highly Cited Paper Related to the Nobel Prize in Physiology or Medicine

作　　者：郭倩影 赵丹群[1] Guo Qianying;Zhao Danqun(Department of Information Management,Peking University,Beijing 100871)

机构地区：[1]北京大学信息管理系,北京100871

出　　处：《情报理论与实践》2024年第10期183-189,182,共8页Information Studies:Theory & Application

摘　　要：[目的/意义]引文主题识别/分析(CTR/CTA)是引文内容分析(CCA)的一项重要研究议题,通过对引文语料中蕴涵主题信息的识别和提取,可望为论文学术贡献评价、知识扩散及演化分析等问题的解决提供新的研究思路。[过程/方法]以一篇高被引诺贝尔生理学或医学获奖关键论文为例,采用BERTopic算法对其引文句语料进行主题识别,并对识别出的引文主题展开多个维度的分析与讨论。[结果/结论]对高被引论文开展引文主题识别分析,有助于更全面细致地揭示其学术贡献内容及演化趋势;BERTopic算法能较好识别案例文献的多个引文主题,且不同引文主题的施引文献特征分布不尽相同;对引文主题重要性、演化趋势及其与原文主题差异性的分析,能多维度刻画研究同行对案例文献学术贡献的认识,表明CTR/CTA研究对学术论文评价具有深入探索价值。[Purpose/significance]Citation Topic Recognition/Analysis(CTR/CTA)is one of the important research fields in Citation Content Analysis(CCA).Identifying/extracting topics from the citation corpus is expected to provide a new solution for the further research on the evaluation of academic contributions,analysis of knowledge diffusion and evolution.[Method/process]This paper takes a highly cited paper related to the Nobel Prize in Physiology or Medicine as an example,uses the BERTopic algorithm to identify its citation topics based on its citation sentence corpus,and then analyzes and discusses the citation topics from multiple dimensions.[Result/conclusion]Our research shows that CTR/CTA of a single highly cited paper can reveal its academic contributions and their evolution trend in a more comprehensive and detailed way.Specifically,BERTopic algorithm can well identify citation topics in our case paper;the characteristics of citing papers vary among different citation topics;the analysis of importance and evolution trend of citation topics and their differences from its original text topics can describe peers’understanding of academic contributions to case paper from different dimensions,which further indicates that CTR/CTA research has in-depth exploration value using for academic paper evaluation.

关 键 词：BERTopic算法 引文主题识别 引文主题分析 引文内容分析 学术论文评价 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]