Sentiment Analysis And Opinion Mining Of Reports On COVID-19 In Chinese And American Mainstream Media

As the COVID-19 pandemic has become a global public health emergency,many mainstream media outlets have continuously reported on the situation in China.However,the evaluations of the Chinese response to the pandemic have been inconsistent across various reports.Therefore,it is essential to compare and understand the sentiment orientation and opinion distribution presented by the Chinese and American mainstream media in their coverage of the COVID-19 outbreak in China.To address this issue,this study utilized a quantitative and qualitative approach based on Liu Bing’s sentiment analysis framework.The collection process used the time nodes of the first through the fourth stage of COVID-19 development in China,as documented in the white paper Fighting Covid-19 China in Action released by the State Council Information Office.The author collected a total of 2408 reports from China Daily and 1041 reports from The New York Times,and constructed a self-built corpus.Natural language processing techniques were employed to calculate the sentiment values of each document using the sentiment lexicon approach.The Sentimentr package and its algorithms were used to perform sentiment analysis and opinion mining at the sentence level,while a topic model was constructed using Latent Dirichlet Allocation to generate different topic distributions of varying sentiments,thereby mining the opinions of the Chinese and American mainstream media on different topics related to the COVID-19 outbreak in China.The major findings of this study are as follows:(1)At the document level,mainstream media in China and the United States exhibited an overall negative sentiment towards the COVID-19 outbreak in China,although the distribution of sentiment varied significantly.American media featured a large amount of negative coverage of the outbreak,with negative sentiment documents accounting for 92.99%of the total.In contrast,negative sentiment documents in Chinese media accounted for50.33%.As the Chinese government achieved significant strategic results in the fourth stage of COVID-19 prevention and control,the sentiment of Chinese mainstream media became more positive and closer to the actual situation,while American mainstream media exhibited a negative sentiment.(2)At the sentence level,the sentiment orientation of mainstream media in China and the United States changed similarly at different stages.Overall,positive,neutral,and negative sentences accounted for 54.24%,10.55%,and 35.21% respectively in Chinese media,while the corresponding percentages in American media were 36.81%,17.63%,and 45.56%,indicating a different overall sentiment orientation.Chinese mainstream media exhibited a positive sentiment orientation at the sentence level,focusing on the positive international evaluations of China’s anti-epidemic efforts and the effective prevention and control measures adopted by the country.American mainstream media exhibited a negative sentiment orientation,criticizing China’s epidemic prevention measures and emphasizing the shortage of medical resources in China.(3)At the aspect level,the main topics of Chinese mainstream media were the impact of the epidemic on the economy,the diplomatic situation under the epidemic,and China’s measures to combat the epidemic.In contrast,the main topics of American mainstream media focused on blaming China for causing the spread of the epidemic,criticizing China’s quarantine measures,and highlighting the shortage of medical resources and the overwhelming burden on the medical system in Wuhan.In conclusion,this study contributes to the research on sentiment analysis and opinion mining in news discourse related to the COVID-19 pandemic.Moreover,it provides insight into the sentiment orientations and opinions of Chinese and American mainstream media towards China’s anti-epidemic measures,thereby facilitating a more objective assessment of news content by the public.