摘要
针对微博短文本提出一种将微博主题与微博情感进行协同分析的方法,利用云南省主流微博媒体数据,研究重大突发公共卫生事件情境下的民众情感极性特征,有助于舆情监测和舆论引导。利用高频词分析研究时间窗口内微博热点主题词,然后训练基于SnowNLP的情感分类模型预测微博情感极性,综合微博信息影响力强度利用隐含狄利克雷分布(LDA)主题模型建模,结合每日疫情实时通报,分析微博情感随时间序列变化趋势。研究发现新冠肺炎疫情爆发以来,多数微博呈现极端正面情感,且微博情感分类结果具有一定的时间聚集性,舆情情感会随公众聚焦事件出现反转,须加以有效引导。
Sentiment analysis in public health emergency is helpful for monitoring and guiding public opinion. The data collected in Yunnan Province of microblogs were used for sentiment analysis and a cooperating method was proposed for analysis of topic and sentiment of messages in microblog in the paper.The Word Cloud was used to analyze hot topic in microblog during research time window,and then sentiment classification modeling based on SnowNLP was trained to obtain microblog sentiment index based on time series. The trend of microblog sentiment with time series was analyzed combining with weighted latent Dirichlet allocation(LDA) topic model and everyday real-time reporting of epidemic situation. The results showed that most of the microblog classification belonged to extreme positive emotions and characterized with time aggregation since the outbreak of Covid-19 pandemic. It was also found that the public opinion and sentiment would reverse with public focus events and positive guidance was necessary.
作者
董婧
范全润
张顺吉
DONG Jing;FAN Quan-run;ZHANG Shun-ji(College of Information Engineering,Qujing Normal University,Qujing 655011,Yunnan,China;Information and Educational Technology Center,Qujing Normal University,Qujing 655011,Yunnan,China)
出处
《内蒙古师范大学学报（自然科学版）》
CAS
2022年第5期489-493,510,共6页
Journal of Inner Mongolia Normal University(Natural Science Edition)
基金
国家自然科学基金资助项目(61841205)
云南省科技厅高校联合面上基金资助项目(2017FH001-056)
云南省教育厅科学研究基金资助项目(2021J0498)
云南省哲学社会科学教育科学规划资助项目(AD17012)。
