摘要
[目的/意义]基于社交媒体用户的共同评论关系构建共评网络,综合运用社会网络分析与自然语言处理技术,探索高效挖掘社交媒体中主流网络民意的方法。[方法/过程]按阶段梳理社交媒体用户共评关系并构建共评网络,综合利用K核分解和核塌缩分析识别核心评论用户群;以核心评论用户群为目标分析对象,从主题和情感两个维度构建主流网络民意的表达,并分析网民讨论热点及情感分布的综合演化过程;利用新冠病毒感染疫情相关热门微博的评论数据进行实证研究。[结果/结论]共评网络分析可以准确识别出社交媒体中的核心评论用户群,其拥有结构稳定且联系紧密的共评关系;聚焦于核心评论用户群的评论内容,即可实现主流网络民意的高效挖掘,准确呈现出网民主要诉求和情感的变化特征;实证结果与我国新冠病毒感染疫情中的应对实情和网络舆论走势基本契合,证明了此方法的有效性。
[Purpose/significance]Based on the common commentary relationships of social media users to build a co-comment network,we use a combination of social network analysis and natural language processing techniques to explore a way to efficiently tap into mainstream online public opinion in social media.[Method/process]Combing social media users’co-comment relationships and constructing co-comment networks by stages,and identifying core comment user groups by using K-core decomposition and core collapse analysis.Using the core comment user groups as the target analysis objects,constructing the expression of mainstream online public opinion from two dimensions:theme and emotion,and analyzing the comprehensive evolution process of hotspots of Internet users’discussion and emotion distribution.Empirical research was conducted using comment data of popular microblogs related to the COVID-19.[Result/conclusion]The analysis of co-comment networks can accurately identify the core comment user groups in social media,who have a stable structure and closely connected co-comment relationships.Focusing on the comment contents of the core comment user groups,it can achieve efficient mining of mainstream online public opinion and accurately present the changing characteristics of the main demands and emotions of netizens.The empirical results basically fit with the response situation and the trend of online public opinion in the COVID-19 in China,which proves the effectiveness of the method.
出处
《情报理论与实践》
CSSCI
北大核心
2023年第10期138-146,137,共10页
Information Studies:Theory & Application
基金
国家自然科学基金项目“信息生态链视角下面向公众的应急信息协同机制与优化策略研究”(项目编号:71974102)
南京邮电大学2022年度引进人才科研启动基金(人文社科类)项目“重大突发事件社会风险的信息发酵机制及治理对策研究”(项目编号:XK1124522041)的成果。