由上海对外经贸大学人工智能与变革管理研究院研究团队成员贾堃（硕士研究生）、祝艺真（硕士研究生）、张钰歆、刘峰*、齐佳音*合作撰写的学术论文“International public opinion analysis of four Olympic games: From 2008 to 2022”被CSCD核心期刊《Journal of Safety Science and Resilience》（《安全科学与韧性（英文）》）录用。该期刊由中国科学院主管，旨在发展为安全科学领域的具有引领性和前瞻性、高度学术信誉与权威度及国际影响力的国际学术期刊，主要发表关于安全科学与复原力实践的科学研究，以及该领域的最新改进。
研究成果基于从Twitter、YouTube等社交媒体平台以及各国传统大众媒体中获取的超过100万条中英文UGC(User Generated Contents)进行情感分析，以比较两种语言在国际舆情中的异同，探讨了上述四届奥运会的国际舆论话题演变和情感差异，为2022年北京冬奥会的国际舆论引导和媒体应对提供了现实的依据。
成果摘要
Abstract:
Since the rapid spread of the COVID-19 worldwide, the pandemic has led to a huge impact on global sporting events. As a major international event, the 2022 Beijing Winter Olympics has commonalities with the 2008 Beijing Olympics, the 2014 Sochi Winter Olympics, and the 2020 Tokyo Olympics in terms of international public opinion context and epidemiological background. In this study, over 1 million pieces of UGC(User Generated Contents) were obtained from the social media platform such as Twitter, YouTube, as well as traditional mass media in various countries.Using sentiment analysis, this study explores the evolution of international public opinion topics and sentiment differences among the above four Olympic Games. The analysis results show that:1) regardless of traditional mass media or online social media, there is a more obvious tendency of general politicization in the topics of the 2008 Beijing Olympics and 2022 Beijing Winter Olympics, and extreme emotional remarks of the 2022 Beijing Winter Olympics are more frequent; 2) in the topic of political opinion involving China, international Chinese public opinion presents more negative sentiment than those in English; 3) among the topics related to COVID-19 involving China, negative public opinion is overwhelming, with huge emotional fluctuations；4)International public opinion on the topic of sports events is significantly more positive in Chinese than in English;5) YouTube's Chinese opinion environment is better than English.
Keywords:
public opinion; sentiment analysis; Olympic Games; COVID-19 pandemic; Computational Affection