图书情报工作 ›› 2021, Vol. 65 ›› Issue (8): 74-84.DOI: 10.13266/j.issn.0252-3116.2021.08.008

• 情报研究 • 上一篇    下一篇

网络突发事件中社交机器人情感的交互式影响机制研究

马晓悦1, 孟啸1, 王镇1, 刘益东2   

  1. 1 西安交通大学新闻与新媒体学院 西安 710049;
    2 上海市浦东微热点大数据研究院 上海 201203
  • 收稿日期:2020-10-16 修回日期:2021-01-11 出版日期:2021-04-20 发布日期:2021-06-02
  • 作者简介:马晓悦(ORCID:0000-0003-4932-6450),特聘研究员,博士,博士生导师,E-mail:xyma_mail@163.com;孟啸(ORCID:0000-0002-7551-5856),硕士研究生;王镇(ORCID:0000-0002-1859-7605),硕士研究生;刘益东(ORCID:0000-0002-0771-6365),学士。
  • 基金资助:
    本文系教育部人文社会科学研究规划基金"信息协同视角下基于可视化媒介的智慧应急响应行为研究"(项目编号:19YJA870009)和陕西省自然科学基础研究计划一般项目-面上项目"基于散射-叠加效应的新媒体信息演化模型构建及事件类别判定研究"(项目编号:2020JM-056)研究成果之一。

Interactive Sentimental Influence Mechanism of Social Bots During an Emergency Event

Ma Xiaoyue1, Meng Xiao1, Wang Zhen1, Liu Yidong2   

  1. 1 School of Journalism and New Media, Xi'an Jiaotong University, Xi'an 710049;
    2 Wei Re Dian Big Data Research Institute, Shanghai 201203
  • Received:2020-10-16 Revised:2021-01-11 Online:2021-04-20 Published:2021-06-02

摘要: [目的/意义] 细化社交机器人对网民情感的干预机制,同时从实践上为网络突发事件舆情治理提供建议参考。[方法/过程] 以仁济医院赵晓菁事件为例,利用朴素贝叶斯方法计算微博情感倾向,通过构建向量自回归模型(VAR)并进行格兰杰因果检验、脉冲响应分析以及方差分解分析,确定社交机器人、意见领袖与普通用户在事件生命周期各阶段的情感关系。[结果/结论] 社交机器人、意见领袖与普通用户的情感关系随舆情阶段演进发生变化,在爆发期,社交机器人放大了意见领袖对普通用户的情感影响;在成熟期,社交机器人影响式微,普通用户的情感反作用于社交机器人与意见领袖;在衰退期,三者保持较为独立的情感关系。此外,社交机器人的影响策略具有隐匿性和间接性特征。

关键词: 突发事件, 社交机器人, 情感分析, 时间序列分析

Abstract: [Purpose/significance] This study further elaborates the intervention mechanism of social bots on netizens' sentiment and provides suggestions for the management of public opinions in online emergencies.[Method/process] Taking the incident of Zhao Xiaojing in Renji Hospital as an example, with the Naive Bayes method calculating Weibo sentiment orientation, this study used the Granger causality test, impulse response analysis, and variance decomposition analysis based on the vector autoregression model (VAR) to judge the sentiment relationship between social bots, opinion leaders and ordinary users at all stages of the event life cycle.[Result/conclusion] The sentiment relationship between social bots, opinion leaders, and ordinary users changed with the evolution of the event. In the outbreak stage, social bots amplified the sentiment influence of opinion leaders on ordinary users. In the mature stage, the influence of social bots declined, and the sentiment of ordinary users reacted to social bots and opinion leaders. In the decline period, the three maintained a relatively independent sentiment relationship; Besides, the influence strategies of social bots were both insidious and indirect.

Key words: emergency, social bots, sentiment analysis, time series analysis

中图分类号: