基于情感分析的网络舆情共振研究

Research on network public opinion resonance based on sentiment analysis

为有效应对复杂多变的舆情环境,研究孤立发生的单一舆情事件演化成具有某些相同特征的多舆情事件簇或事件集,针对同议题、同情绪等多起舆情事件建立网络舆情共振模型,通过爬取“唐山打人案”和“唐山打人事件被害人首次发声”事件相关微博数据,以及“2021年河南遭遇特大暴雨”和“2023年河北暴雨”相关微博数据,将评论数据进行BosonNLP情感分析,得出其情感分数;并将情感分数作为模型参数,分别对2起不同类型的案例进行检验。研究结果表明:在原生舆情与次生舆情共同作用下,引起网民情绪感染和矛盾冲突,从而发生网络舆情共振,并且共振产生的热度高于单一事件热度;网民消极的态度值会加速舆情共振、不同类型的事件所产生的舆情共振效果是不同的。研究结果可丰富网络舆情以及社会物理学相关理论,可为构建舆情共振的研究框架提供参考。

In order to cope with the complex and changing public opinion environment,the evolution of single public opinion event occurring in isolation into a cluster or a set of multiple public opinion events owing certain same characteristics was studied,and a resonance model of network public opinion for multiple public opinion events with the same topic and same emotion was established.Through crawling the Sina-Weibo data including 2022 Tangshan beating incident and the victim’s first voice,as well as the torrential rain in Henan in 2021 and the torrential rain in Hebei in 2023,the BosonNLP sentiment analysis of the comment data was conducted to obtain its sentiment score.By utilizing the sentiment score as the model parameter,two different types of cases were tested respectively.The results show that the primary public opinion events and secondary public opinion events work together to cause the netizens’emotional infection and conflict,thus the network public opinion resonance occurs,and the heat generated by the resonance is much higher than that of single event.The netizen’s attitude value,as well as the relevance of issue index,and the value of the event’s heat,are the keys to judge whether the resonance occurs or not.The research results can enrich the theories of network public opinion and social physics,and provide a reference for constructing a research framework of public opinion resonance.