基于社会化媒体的谣言受众反应分析方法及干预策略研究

At present, rumours on social media sites occur frequently in China, which often have an enormous impact on public life, business operation and social order. However, rumours themselves are not destructive. It is the subsequent collective reactions of the audience that generate great influence and destructiveness. As such, based on the data collected from social media sites,this project investigates the analytical approaches and intervention strategies of the audience's reactions to rumours, including (1) to propose methods to measure sentiment orientations of the individual and the group who are affected by a rumour, analyze internal and external factors influencing the group sentiment and its changing pattern, and study intervention strategies that can lead the group sentiment to the desirable orientation; (2) to present methods of creating and analyzing semantic network based on audience generated content related to a rumour, analyze the audience's underlying appeals and the reasons behind them, and study intervention strategies that can guide the audience to express more reasonable appeals; and (3) to propose a method to calculate to what degree an individual's opinion is in accordance with others, study the method of identifying the influencial people based on the measurement of the degree of accordance, analyze the influence mechanism of influencial people on the group's interactive network evolution, sentiment changes and appeal changes, and finally propose intervention strategies to influencial people.This project is expected to contribute to the literiture of rumours on social media sites and help the management of virtual society in aspects of theory and methods.

目前中国社会化媒体上谣言频发，常常对公众生活、企业运营和社会秩序带来严重影响。然而谣言本身并不具有破坏性，受众面对谣言时的趋同反应才是其产生巨大影响力和破坏力的主要因素。本课题基于社会化媒体上的天然数据,研究谣言受众反应的分析方法及干预策略：①提出受众个体反应情感倾向和受众群体情感的测量方法，分析受众群体情感的内外部影响因素及变迁规律，探讨引导受众情感朝良性状态迁移的干预策略；②探索受众反应语义网络的构建方法和分析方法，挖掘受众反应诉求及其背后的共同因子，并提出引导受众诉求的干预策略；③研究受众观点被其他受众认同程度的计算方法，提出基于观点认同度的有影响力受众的识别方法，分析有影响力受众对受众群体交互网络演化、情感倾向变迁和诉求形成的影响机制，并探讨针对有影响力受众的干预策略。本课题将推动基于社会化媒体的谣言理论研究进展，为相关部门提高虚拟社会管理水平提供理论基础和方法支持。

本项目采用真实网络谣言和网络口碑数据，基于Penal data模型、VAR模型、HLM多层线性模型等多种计量模型，主要开展了网络谣言的识别研究、受众对网络谣言的关注度研究、网络口碑与管理者反馈(干预策略)的商业价值研究以及有影响力受众对受众情感倾向变迁的影响机制研究等内容，在这些重点研究内容上取得了一系列的研究成果，圆满地完成了项目申请时的预期成果(学术论文9篇，其中至少3篇为SCI/SSCI检索)，项目批准立项三年来，共发表期刊论文12篇，其中SSCI/SCI期刊论文8篇、EI期刊论文1篇、CSSCI期刊论文2篇、国际期刊论文1篇。