The Current Situation Of Internet Public Opinion Governance In COVID-19

With the rapid development of financial media,the channels for the public to obtain information are becoming more and more abundant,and they express some personal comments and opinions on the network.The official mainstream media also quickly adapted to the development of the situation and carried out self-innovation and transformation.At the same time,we media also entered a stage of rapid development.When the epidemic comes,the official mainstream media,in addition to conducting public opinion guidance on traditional information dissemination channels,also conducts multi angle,high-frequency and active online public opinion guidance on the new financial media platform,but at the same time,some problems also arise in this sudden situation.Based on governance theory,emotional sociology theory,etc.,this paper analyzes the evolution characteristics of online public opinion in the COVID-19 incident by using a variety of statistical models,including K-Means clustering algorithm,CURE algorithm,text representation model,etc.Through Steven Fink’s four stages of crisis communication and their characteristics,the following points are studied:First of all,this paper proposes a construction method of emotion dictionary based on context and semantic information,which maximizes the coverage of emotional words in the model.Secondly,the text model is used to realize the corresponding relationship between theme and emotion.Finally,the k-means clustering method is improved to automatically obtain the initial basic cluster center.Through the improved statistical model,the corresponding research results are obtained,and some suggestions for network public opinion guidance are put forward.