Research On Early Warning Of Public Opinion Crisis Under Multiple Scenarios Of Public Emergencies

At present,China’s comprehensive deepening of reform has entered a period of tackling difficult problems,the domestic development environment has undergone profound changes,and various social contradictions have become increasingly prominent.In particular,the frequent occurrence of public emergencies has become a difficult problem for social governance,which not only has a negative impact on public life,economic development and social harmony and stability,but also triggers a crisis of online public opinion and seriously disrupts the online public opinion environment.Therefore,exploring the early warning process and crisis response strategy of online public opinion crisis of such incidents can help relevant departments scientifically understand the development trend of public opinion,improve crisis handling capabilities,and create a good public opinion environment.In this regard,this paper first sorts out the references at home and abroad from three aspects: early warning of network public opinion crisis,governance of network public opinion crisis and scenario deduction,summarizes the characteristics of scenarios,public opinion crisis warning of public emergencies and public opinion communication of public emergencies,and analyzes the problems existing in the research on public opinion crisis early warning.Secondly,the scenario unit of the event is determined,the scenario is defined as the initial scenario,the development scenario and the end scenario,and the network public opinion crisis early warning scenario is structurally expressed from four aspects of scenario event,scenario driving element,scenario reasoning element and scenario target element based on the knowledge element model,and the dynamic Bayesian network of public opinion crisis early warning scenario is constructed,the event public opinion information is collected,75 event samples are retained,and the dynamic Bayesian network model is trained and tested by EM algorithm.The public opinion crisis state under the three scenarios and the main factors affecting the early warning level are given.Then,the alternative options that can be adopted by the government,the media,opinion leaders and think tanks are analyzed,the index evaluation principle is determined,the optimal evaluation index system of the alternative scheme for early warning response to public opinion crisis is constructed,the subjective judgment factors of experts are considered,the analytic hierarchy method is used to determine the index weight,the projection method is used to determine the expert weight,and the alternative optimal evaluation model is constructed based on the calculation steps of the intuitive fuzzy integral operator.Finally,taking the epidemic incident at Nanjing Lukou Airport as an example,the timeline of the incident is sorted out,the initial scenario,development scenario and end scenario of the event are divided,the driving factors of the scenario are judged,and they are input into the dynamic Bayesian network of the public opinion crisis early warning scenario as evidence variables,the possibility of the outbreak of public opinion crisis under different scenarios is evaluated,and three experts in the field of public opinion research are invited to evaluate and optimize the alternatives adopted by the government,media,opinion leaders and think tanks in different scenarios.Deriving the best response plan of each entity has certain guiding significance for achieving efficient governance.