Evolutionary Game Study On Public Opinion Of Emergencies Driven By Social Media User Data

Public emergencies refer to natural disasters,accident disasters,public health events and social security events that occur suddenly and cause or may cause serious social harm and need to be dealt with by taking emergency response measures.The emergencies,uncertainties,commonalities,group hazards,dynamics of major public emergencies Complexity not only poses a very serious threat to people’s life safety,causes instability to the society,but also poses a severe challenge to the government’s emergency management.Since the 18 th National Congress of the Communist Party of China(CPC),General Secretary Xi Jinping has repeatedly proposed to modernize national governance with modern technologies,promote innovation in government management and social governance models,and achieve the goals of scientific government decisionmaking,accurate social governance,efficient public services,and intelligent management process.Therefore,improving the capacity-building to deal with major public emergencies is not only a severe challenge that national security governance must face,but also a hot and difficult issue concerned by the current academic circles.With the rapid popularization of mobile devices,the speed of public access to and dissemination of information has accelerated.At the same time,the rapid development of social media platforms has become the promoter of information dissemination and the driving force of the formation of public opinions.On the other hand,the rapidity,explosiveness and uncontrollability of social media platform information release also lead to the frequent occurrence of network public opinion problems.In the context of public emergencies,all kinds of user data information transmitted by social media platforms are highly representative for grasping public behavior and mentality.It is very important to quickly find out the behavior law of social media users,perceive their emotional changes,master the evolution law of public opinion and control the development direction of public opinion.Focusing on the main line of "Data analysis-Algorithm integration-Case study-Simulation",this paper studies a series of problems such as the behavior law of social media users,emotional change and public opinion evolution game under the background of public emergencies: firstly,this paper analyzes the basic characteristics of social media users’ emotional data under the background of public emergencies,A new emotion classification method integrating general knowledge and domain specific knowledge is proposed and its effectiveness is verified.On this basis,through the introduction of actual cases,this paper studies the behavior and emotion evolution process and differentiation of social media users under the background of public emergencies.Based on the conclusion of case study,this paper explores the common Internet users The evolution game process of network public opinion in public emergencies with the participation of ordinary media and government media,draw corresponding conclusions and put forward corresponding policy suggestions.The main work is as follows:(1)Feature analysis and preprocessing of social media basic data under public emergencies.In view of the obvious class imbalance of social media data in the context of public emergencies,a new emotion classification method integrating general knowledge and domain specific knowledge is studied and designed.Different from the existing research on generating dictionary based features from traditional emotional dictionaries,label propagation method is developed to induce corpus adaptive emotional dictionaries.Specifically,we use glove embedding to implement the proposed label propagation algorithm to generate a domain insensitive emotion dictionary,and use SVD embedding in our label propagation model to summarize a domain specific emotion dictionary.The generalized general vocabulary,domain specific vocabulary and the two words are embedded and further fused to train the integrated classifier.In addition,aiming at the unbalanced data involved in emotion classification,a new oversampling method is proposed in this paper.Since the basic unit of text data is words,and the previous oversampling method is mainly carried out in digital space,the proposed oversampling method tries to directly generate artificial text through two text generation strategies: "text substitution of synonyms" and "antonyms".Experiments show that the algorithm is reliable,interpretable and effective in the emotional classification of unbalanced text data.(2)Research on the behavior and emotional evolution of social media users under public emergencies.Through the topic tag function of microblog platform,this paper analyzes the data of different topics generated by social media platform under the same event,establishes a framework to track the evolution of social media user behavior and public emotion,and analyzes the correlation between them.The study found that in the context of public emergencies,the behavior of social media users is basically consistent with their normal work and rest,and social media users have not deliberately changed their schedule because of public emergencies.With the development of events,new hot topics came into being,quickly attracted the attention of users,and led to the decline of people’s interest in the previous topics.However,the change of discussion heat of different hot topics under the same public emergency has a similar trend.Through the research on users’ concerns,it is found that the most important function of sina Weibo for users is to express their feelings and emotions.Users’ emotions vary greatly under different hot topics.Users’ emotions are positively related to the information released by individual authenticated users and the maximum amount of information per minute.This study provides a new perspective for information extraction from social media platforms and the law of emotional communication in public emergencies.(3)Research on social media user behavior and emotional differentiation under public emergencies.Based on the results of the previous part,this part of the research is mainly devoted to analyzing the differences between different social media users in behavior law,emotional change and influence under the background of public emergencies,and on this basis,it analyzes the differences of users on social media platforms at home and abroad.The research results are intended to provide a basis for government agencies to deal with similar trust crises.First of all,when a crisis occurs,government agencies must communicate timely and effectively.Research shows that when a public emergency occurs,social media users tend to obtain information from users authenticated by the social media platform.Similarly,government agencies can maintain some continuity on their social media platforms to effectively evaluate and maintain public sentiment in the event of public emergencies.This study provides a new idea for government agencies to develop platform authentication users,and provides a new idea for government agencies to reasonably use platform authentication users in the event of trust crisis.(4)Research on the evolution game of social media platform network public opinion under public emergencies.Based on the research results of the first two parts,this study aims to explore the evolutionary game process of network public opinion in public emergencies with the participation of ordinary Internet users,ordinary media and government media.This part analyzes the impact of the effectiveness of government media intervention in information dissemination on the evolution of public opinion,and how to improve the effectiveness of government media intervention in information,so as to better build a healthy and harmonious network environment and maintain social order.The research constructs the payment matrix and the replication dynamic equation of the three parties according to the parameter hypothesis,and completes the construction of the three-party evolutionary game model.The results show that the effectiveness of government media intervention will affect the strategic choice of the three parties,thus affecting the results of the three-party evolutionary game and the evolution of network public opinion in public emergencies.The government media need to fully grasp the network public opinion information,not to allow the evolution of the network public opinion of public emergencies excessively,but also need to grasp the network rumor information,refute the rumor,intervene the network public opinion of public emergencies in time,and build a healthy and harmonious network environment.The purpose of this study is to explore the behavior law,emotional change and public opinion evolution law of social media users under public emergencies.Comprehensively consider the development stage of public opinion,obtain the emotion and behavior related data in news reports,microblogs and other media through data mining technology,master the research of the whole process of behavior law,comprehensively consider the influence mechanism of events,the public,the media and the government on the evolution of public opinion,establish a complex system model,master the evolution law,and guide emotional behavior in order to control the development of public opinion,The government takes effective measures to provide theoretical basis.The main innovations are as follows:(1)Aiming at the problem of domain sensitivity and imbalance in emotion classification of text data,an emotion classification model integrating general knowledge and domain specific knowledge is proposed based on oversampling technology.(2)In view of the existing research on the evolution law of public opinion,which does not fully consider the comprehensive impact of different factors in different stages on public sentiment,a multi-dimensional dynamic social media user behavior and public sentiment evolution analysis framework is constructed.(3)Aiming at the limitation that most of the existing researches on public opinion intervention are single factor,based on the interactive relationship between ordinary Internet users,ordinary media and government media,this paper constructs a three-party subject dynamic game model.