Journal of Data Science | |
On Public Sentiment and Topic Mining during the COVID-19 Pandemic Based on Sina Weibo Comment Data | |
article | |
Xiaomeng Du1  Wei Huang1  Yijing Liu1  Haibo Su1  | |
[1] Data Science Lab | |
关键词: Dirichlet distribution; emotional classification; naive Bayes; visualization; word co-occurrence; | |
DOI : 10.6339/JDS.202012_18(5).0003 | |
学科分类:土木及结构工程学 | |
来源: JDS | |
【 摘 要 】
In the wake of the COVID-19 outbreak, the public resorted to Sina Weibo as a major platform for the trend of the pandemic. Research on public sentiment and topic mining of major public sentiment events based on Sina Weibo’s comment data is important for understanding the trend of public opinions during major epidemic outbreaks. Based on classification of the Chinese language into emotion categories in psychology, we use open source tools to build naive Bayesian models to classify Weibo comments. Visualization of comment topics is achieved with word co-occurrence network methods. Commented topics are mined with the help of the latent Dirichlet distribution model. The results show that the psychological sentiment classification combined with the naive Bayesian model can reflect the evolvement of public sentiment during the epidemic, and that the latent Dirichlet distribution model and word co-occurrence network can effectively mine the topics of public concerns.
【 授权许可】
CC BY
【 预 览 】
Files | Size | Format | View |
---|---|---|---|
RO202307150000425ZK.pdf | 15276KB | download |