| With the development and popularity of the Internet,social networks have gradually become a channel for people to exchange information.Internet users publish various opinions through social networks,and thus online public opinion is created and developed.Online public opinion often leads to a series of negative situations,including the generation of mass panic and influencing the real behavior of Internet users.Therefore,effective management of online public opinion becomes the key to establishing a good online order.Especially for online public opinion generated by unexpected events,its suddenness and rapid growth make it more difficult to control.Scholars in the field of public opinion research often analyze the evolutionary characteristics of public opinion by establishing public opinion dynamics models.It is of great theoretical significance to establish an effective public opinion dynamics model to guide relevant departments to control public opinion on specific emergencies according to the characteristics of such events.In this paper,based on examining a large amount of relevant literature and combining relevant theories and literature,we firstly propose an improved Deffuant model based on the classical Deffuant public opinion dynamics model in four dimensions:interaction environment,initial opinion distribution,group heterogeneity and influence of other subjects,and establish an improved public opinion dynamics model that is more in line with the changing characteristics of public opinion of an emergency event.Meanwhile,python programming is used to design the corresponding algorithms as well as simulation experiments for the improved model from the above four aspects.Then,the crawler method was used to crawl the public opinion data of the sudden natural disaster "He Nan rainstorm",and the analysis of public opinion characteristics was carried out in stages and groups,based on which the sentiment analysis method was used to analyze the crawled public opinion data to obtain the opinion values,and the opinion values were substituted into the improved model.The rationality and validity of the improved model are verified through experiments.Finally,based on the new features of public opinion presented by the improved model,the government’s countermeasures and suggestions for coping with public opinion of sudden natural disasters are proposed from three perspectives: the life cycle of public opinion outbreak,the crowd user profile of public opinion,and the government’s governance ability,in order to provide theoretical guidance for the government to cope with public opinion of complex sudden natural disasters.The study shows that for the improved Deffuant public opinion dynamics model,firstly,in terms of interaction environment,the characteristics of BA scale-free network in complex networks are more consistent with the characteristics of realistic public opinion interaction environment.Secondly,in terms of the initial opinion distribution,the distribution of opinion values under the right skewed distribution is more consistent with the characteristics of emergent events.Third,in terms of group heterogeneity,the convergence of the model with the addition of individual relationship strength rules is affected to some extent but is more consistent with the characteristics of group heterogeneity in real public opinion.Therefore,the model with the inclusion of BA scale-free network,right skewed distribution,and inter-individual relationship strength rules is beneficial to the government departments for public opinion warning of sudden natural disaster events.Fourth,when considering the influence of other subjects,opinion leaders and the government will guide the direction of public opinion;the greater the government influence factor,the greater the degree of influence on public opinion,and the greater the government trust,the greater the influence on public opinion;therefore,the government’s public opinion governance capability should be strengthened by strengthening the government’s influence on public opinion,focusing on enhancing the government’s credibility,and strengthening the interaction with opinion leaders.In addition,based on the characteristics of public opinion embodied in cases of sudden natural disasters,it is important to pay attention to the evolution direction of public opinion during the outbreak period,to guide public opinion during the recession period and to monitor public opinion during the stability period.For groups of different age,gender and geographical characteristics,it is necessary to grasp the characteristics of each group for targeted guidance. |