A Study On How Twitter's Socialbots Interfere Public Opinion Related To China During Early Coronavirus Epidemic

With the development of Internet technology,Socialbots have become a common communication force participating in all kinds of public events on social media platform.Since the spread of COVID-19,Twitter users around the world have been using this social media platform to discuss about this epidemic.There is also a lot of interference from Socialbots to public opinion.In order to discuss the interference of Socialbots on public opinion,this study uses data collection software to collect the user's tweets on the Twitter platform as the analysis object,uses the sentiment analysis tool to analyze the characteristics of Socialbots' tweets,and innovatively uses the frame theory to investigate the similarities and differences between Socialbots and human users in the frame selection.After collecting 170529 tweets and measuring their social robot scores,this study found that 15.71% of tweets related to China were published by Socialbots,and the proportion of Socialbots accounted was 17.63%,with an average of 3.2 tweets per Socialbot.Socialbots interfere public opinion with the characteristics of high frequency,high repetition rate and weak originality.More than half of Socialbots accounts use external links while posting tweets,while less than 30% of human accounts use external links.In terms of emotional tendency,there is no significant difference between Socialbots and human users.However,at the important time related to the epidemic,the negative emotions of Socialbots increase significantly,so the interference to public opinion are more negative than other times.The study found that Socialbots and human users pay different attention to issues related to China.Socialbots pay more attention to political and humanistic care issue,while human users pay more attention to human rights and anti-epidemic measure issues.In terms of frame selection,Socialbots have different frame selection strategies for different issues.They mostly use responsibility attribution frame for political issues,moral frame for human rights issues and factual frame for epidemic growth issues.It also found that Socialbots with strong negative emotions and human users with strong negative emotions have significant differences in frame selection,while Socialbots with positive emotions and human users with positive emotions have no significant differences in frame selection.