基于动态贝叶斯网络的突发事件网络谣言风险预警研究

Research on Risk Early Warning of Emergency Network Rumors Based on Dynamic Bayesian Network

[目的/意义]利用动态贝叶斯网络构建突发事件网络谣言风险预警模型,丰富突发事件网络谣言的研究内容,对我国网络谣言治理提供参考。[方法/过程]从事实维度、社会维度和时间维度3个维度确定突发事件网络谣言风险影响因素和量化方法,构建突发事件网络谣言风险等级预测指标体系。随后,构建动态贝叶斯网络并进行训练,以“唐山烧烤店被打女子被车碾压已死亡”事件为例,利用动态贝叶斯网络对其属性进行分析。[结果/结论]社会维度对于风险水平的贡献度最大,时间维度在不同阶段的风险级别有所差异,政府部门应与意见领袖合作,提升辟谣效率。

[Purpose/significance]Using Dynamic Bayesian Network(DBN)to construct the risk early warning model of emergency network rumors,the paper enriches the research content of emergency network rumors and provides reference for the governance of network rumors in China.[Method/process]From the three dimensions of fact dimension,society dimension and time dimension,this paper determines the influencing factors and quantitative methods of emergency network rumor risk,and constructs the prediction index system of emergency network rumor risk level.Then,the paper constructs a DBN and trains it,taking the incident of“Tangshan barbecue restaurant was hit by a woman who was crushed by a car and died”as an example,and using the DBN to analyze its attributes.[Result/conclusion]The social dimension contributes the greatest contribution to the risk level,and the time dimension has different risk level at different stages,and government departments should cooperate with opinion leaders to improve the efficiency of rumor refutation.