Crisis Communication Management Research Of Electronic Word Of Mouth On Social Media

With the rapid development of Internet technology and the continuous popularization of social media technology,social media electronic word of mouth(e WOM)is playing an increasingly important role for enterprises.Meanwhile,in the highly transparent information age,corporate crises,large and small,are exposed to the public eye,which make corporate crises frequent and complex.Compared to other platforms,the social attributes,decentralization and other unique attributes of social media platforms make the enterprise crisis expose faster and spread with the spread of electronic word of mouth on social media.The impact on enterprise is also greater.How to explore the crisis elements in the e WOM communication of social media and achieve reasonable control is of great significance to the enterprise.Among the existing researches on e WOM of social media both at home and abroad,some research results on the mining,analysis and control of e WOM of enterprise crisis have laid a foundation for the followup related researches,but no complete theoretical system has been formed.Based on this,this thesis combines the crisis management theory,natural language processing,machine learning,deep learning,system identification and timing technology methods to study the crisis communication rule and management of e WOM on social media.The research includes:First,this thesis constructs the theoretical framework of crisis communication management of social media e WOM.First of all,the research object and content are defined based on the sorting out of existing theoretical studies such as e WOM on social media.Secondly,combining with relevant theories of crisis communication management,this thesis analyzes the crisis communication process of social media e WOM from the perspective of whole-process management.On this basis,this thesis puts forward a research framework of e WOM crisis communication management on social media,which is centered on text content mining,and includes research contents such as e WOM communication law,disinformation channel mining and crisis communication management methods.Second,from the perspective of crisis regulation,this thesis explores and analyzes the daily communication process of e WOM text on social media.Based on the natural language processing method,text content is collected and preprocessed for text feature expression and text calculation,and communication features are mined and calculated by using semi-automatic dictionary method in combination with marketing theory.Based on the results,from the perspective of crisis regulation,this thesis uses the time-series econometric model method to model and analyze the daily communication process,and reveals the different effects of each characteristic on e WOM.Finally,combined with the practical application of enterprises,this thesis constructs the normal monitoring content of electronic word-of-mouth.Thirdly,this thesis conducts modeling analysis on the e WOM communication process under crisis events,and explores the different influences of different crisis event attribute characteristics on the communication process.First of all,the second order propagation process model under the e WOM crisis of social media is constructed by using methods of system identification time domain analysis.Next,this thesis conducts a quantitative division method for the communication process based on the development law of crisis events,and explores the influence of the important characteristics of the evolution process of crisis events on the e WOM communication process and its final communication volume.Finally,based on the analysis,this thesis puts forward a preliminary prediction method for the communication process of e WOM for enterprise managers according to the crisis attribute information,which provides method support for the establishment of crisis management strategies for enterprise managers.Fourthly,focusing on the malicious behavior in the spread of e WOM crisis on social media,this thesis puts forward a method to identify the disinformation communication.Firstly,this thesis first points out the malicious behavior in the crisis communication of e WOM.Relevant variables describing the characteristics of disinformation and negative information characteristic variables are proposed as characteristic variables to identify the channels of disinformation channel.Based on the actual transmission data,the study verifies the significant positive influence of different vatiables on determining whether the channel is a malicious transmission channel.Finally,based on the machine learning model,the importance of these variables for the identification of disinformation channels is revealed,and an automatic identification model of disinformation channels is constructed,so as to provide decision support for enterprises to control disinformation channels.Finally,the thesis puts forward the management method for the crisis communication of social media e WOM.First of all,based on statistical methods,this thesis explores the influence of different enterprise response modes,response time and response content on the e WOM communication process of social media under crisis events,and proposes reasonable response methods for enterprises in crisis events.Secondly,for the practical application of enterprises this thesis builds the crisis communication monitoring and management process of enterprise social media e WOM,e WOM communication management process and e WOM data management dimension in crisis events.This thesis establishes the management process of e WOM crisis communication in social media,and proposes a direct application scheme for enterprises to better implement e WOM crisis communication management.