突发事件社交媒体用户观点变化研究——基于多层级微博评论视角  

Changing User Perspectives on Social Media During Sudden Events:Based on Multilevel Weibo Comments

作　　者：胡媛[1,2] 廖文涛 艾欣怡 刘婷 HU Yuan;LIAO WenTao;AI XinYi;LIU Ting(School of Public Policy and Administration,Nanchang University,Nanchang 330031,P.R.China;Digital Literacy and Skills Enhancement Center,Jiangxi Key Research Base of Philosophy and Social Sciences,Nanchang 330031,P.R.China)

机构地区：[1]南昌大学公共政策与管理学院,南昌330031 [2]江西省哲学社会科学重点研究基地数字素养与技能提升中心,南昌330031

出　　处：《数字图书馆论坛》2024年第8期29-38,共10页Digital Library Forum

基　　金：国家社会科学基金青年项目“突发公共事件中的网络社会心态演化及调节机制研究”(编号:22CTQ022);江西省高校人文社科项目“突发事件社交媒体用户负面情感早期识别与风险预警机制研究”(编号:TQ22201)资助。

摘　　要：突发事件背景下不同层级的评论可以细粒度地揭示用户观点变化,对捕捉舆论矛盾和网络舆情治理有重要意义。以突发事件的微博评论为研究对象,利用基于TF-IDF算法的LDA模型、情感倾向分析和Kruskal-Wallis差异性检验,从用户关注焦点和情感异同两个方面对比不同层级评论的用户观点变化。研究发现:就用户关注焦点而言,随着评论层级的增加,用户关注焦点发生变化,讨论话题逐渐延伸,一级评论对于事件的针对性、关联性最强,二级评论则更关注事件的发展与处理进程,三级评论中用户间矛盾冲突最为突显;就情感而言,评论均表现出强烈的消极倾向,但各层级评论的用户情感存在显著差异,且评论层级越深,情感越消极,三级评论比一级、二级评论带有更多隐性消极情感用词,符合认知负荷理论。研究为社交媒体用户信息行为和网络舆情分析提供了更加细粒度的视角。In the context of sudden events,comments at different levels can reveal the nuanced changes in user perspectives,which is of significant importance for capturing the contradictions in public opinion and managing online sentiments.This study focuses on Weibo comments related to sudden events and employs a TF-IDF-based LDA model,sentiment analysis,and Kruskal-Wallis variance testing to compare the changes in user perspectives across different comment levels,considering both user focal points and emotional variations.The research findings indicate that,in terms of user focus,as the level of comments increases,the focal points of users change and the topics of discussion evolve.First-level comments exhibit the strongest relevance and specificity regarding the events,and second-level comments pay more attention to the development and management process of the events,while third-level comments highlight the conflicts and contradictions among users.Emotionally,all levels of comments express a strong negative tendency,but significant differences in emotional expressions exist across the levels:deeper levels exhibit increasingly negative sentiments,with third-level comments containing more implicit negative emotional language than first-level and second-level comments,aligning with cognitive load theory.This study provides a more specific research perspective on the information behavior of social media users and online sentiments.

关 键 词：突发事件 社交媒体 多层级评论 用户观点 认知负荷理论 

分 类 号：G206[文化科学—传播学]