【摘要】 从信息化语境中挖掘社交媒体平台中的舆情数据，尝试更加系统地分析新冠疫苗接种意愿多重影响因素间的关系.利用网络爬虫技术抓爬社交媒体Twitter平台中的数据，借助潜在狄利克雷分配(latent Dirichlet allocation, LDA)模型识别影响新冠疫苗接种意愿的因素，并运用解释结构模型方法(interpretative structural modeling method, ISM)辅助建模，构建多级层次递阶结构模型，梳理和分析各项影响因素之间的逻辑层次关系.结果表明：通过LDA和ISM模型可以对获取到的因素集进行结构化处理，并根据ISM模型的逻辑结构，将影响新冠疫苗接种意愿的4个层级划分为深层、中间层和表层，不同层级对疫苗接种意愿的影响程度不同.研究结果可为提升民众对新冠疫苗的接受意愿和接种倾向提供相应的决策支持和参考价值.更多还原

【Abstract】 Mining public opinion data from the social media platform in the information context, we try to more systematically analyze the relationship between multiple influencing factors of COVID-19 vaccination willingness. Focusing on the research problems of influencing factors of the vaccination intention of the COVID-19 vaccination, we use the web crawler technology to crawl the data in the social media Twitter platform, mine the main factors affecting the vaccination intention of the new coronal vaccine with the help of the Latent Dirichlet Allocation theme model, and use the Interpretive Structural Modeling Method to assist modeling, build a multi-level hierarchical structure model, so as to sort out and analyze the logical hierarchical relationship between the influencing factors. The results show that the factor set can be structurally processed through LDA and ISM models, and according to the logical structure of the ISM model, the four levels that affect the willingness to vaccinate COVID-19 vaccine are divided into the deep layer, the middle layer and the surface layer. Different levels have different degrees of influence on the willingness to vaccinate. The research results can provide corresponding decision support and reference value for improving people’s willingness to accept COVID-19 vaccine and vaccination tendency.更多还原