| 摘 要: 针对社交媒体上虚假信息传播能力影响因素不明确的问题,提出了一种结合 LDA 的主题建模和模糊集定性比较分析(fsQCA)的潜在狄利克雷分配研究方法。一方面,运用 LDA 技术构建微博用户评论的主题分析图谱,提取出6个关键主题。另一方面,基于刺激-有机体-反应(S-O-R)框架,以6个主题为条件变量,以传播能力为结果变量进行了fsQCA分析,探究影响虚假信息传播能力的组态路径。实验结果表明,单一主题对高传播能力虚假信息的形成影响有限。通过对组态路径的分析,对虚假信息的高/低传播能力各归纳了两类形成机制。关键词:虚假信息传播;LDA;主题建模;模糊集定性比较分析;S-O-R框架 |
| 关键词: 虚假信息传播 LDA 主题建模 模糊集定性比较分析 S-O-R框架 |
|
中图分类号: TP39
文献标识码: A
|
|
| Research on the Configuration Analysis and Formation Mechanisms of Misinformation Dissemination Capability Basedon LDA |
|
WANG Shixiong, QI Qi
|
(School of Economics and Management, Zhejiang Sc-i Tech University, Hangzhou 310018, China)
sxwang@zstu.edu.cn; qiqi_zstu@outlook.com
|
| Abstract: In response to the unclear influencing factors of misinformation dissemination capability on social media, this study proposes a research method that combines LDA(Latent Dirichlet Allocation)topic modeling and fsQCA(fuzzy-set Qualitative Comparative Analysis) configuration analysis. On one hand, LDA technology is employed to construct a thematic analysis map of Weibo user comments, extracting six key themes. On the other hand, based on the S-O-R(Stimulus-Organism-Response)framework, with the six themes as conditional variables and dissemination capability as the outcome variable, an fsQCA analysis was conducted to explore the configuration paths affecting misinformation dissemination capability. Experimental results indicate that individual themes have a limited impact on the formation of highly disseminated misinformation. Through the analysis of configuration paths, two types of
formation mechanisms were summarized for both high and low dissemination capabilities of misinformation. |
| Keywords: misinformation dissemination LDA (Latent Dirichlet Allocation ) topic modeling fuzzy-set Qualitative Comparative Analysis (fsQCA) S-O-R framework |
