Analyzing Emergency News Comments to Assess Online Public Opinions: A Case Study of COVID-19

Abstract

For the purpose of enhancing the online public opinion intervention mechanism, fostering a positive public opinion environment, it is crucial to examine the rules and traits of online public opinion dissemination from multiple perspectives. The paper proposes a fresh approach to measuring network public opinion by analyzing emergency news comments. In order to develop risk assessment indicators, we first employ the multi-round Delphi method. Then, we organize the "COVID-19" news comments using natural language processing and text clustering techniques, correlating the risk assessment indicators with the risk evolution of emergency events. Finally, we analyze the time evolution trend of users’ participation in network public opinion. Results show that the use of news commentary can effectively predict the tendency of social risk. Therefore, the risk assessment method in this paper can judge and warn the network public opinion in time, which is of great value in assisting major emergency management decisions.