International Public Opinion Analysis in COVID-19 Pandemic from Mainstream and Social Media

COVID-19 has become a global pandemic, and it is crucial to study and assess international public opinion during this period. In this study, we have chosen China as a typical case to analyze international public opinion during COVID-19. Our research aims to address the following questions: (1) What is the correlation between the results of public opinion analysis and factors such as polarity, the density of active words, and the density of pronouns? (2) What aspects of China do foreign media focus on during the COVID-19 period, and what are their public opinion analysis results in these areas? (3) Does the severity of the epidemic affect the degree of concern of foreign media on China? To conduct our analysis, we collected 735,566 pieces of data from GDELT and millions of pieces of data from Twitter as the object of our analysis. We combined adversarial learning for zero-shot sentiment analysis and PLS (Partial least squares regression) to propose a new method for public opinion analysis. Our model analysis revealed that the results of public opinion analysis were related to several factors, including polarity, active words, and the proportion of pronouns. Among these factors, polarity had the most significant influence.