[目的/意义]研究旨在定量测量突发公共卫生事件多元主体的信息披露水平，为研究突发公共卫生事件
多元主体的信息披露行为提供新的分析框架。从数据层面揭示我国疫情期间信息披露的现状，为政府等相关主
体调整信息发布内容、改善信息披露行为提供理论支撑。[方法/过程]以各主体发布的疫情相关微博为研究对象，
构建多元主体信息披露内容及特征演化分析框架。利用改进的 TFIDF-LDA模型进行主题提取，基于社会渗透理
论测量社交媒体多元主体信息披露的广度和深度。最后结合生命周期理论分析信息披露内容特征，探究多元主
体信息披露特征及演化趋势。[结果/结论]疫情期间多元主体的信息披露内容主题广泛，不同主体发布内容偏好
和披露特征存在明显差异。政府的信息披露内容积极、质量最高，旨在正向引导舆论；主流媒体的信息披露范围
广深度浅，对数据隐私保持高度敏感；自媒体的信息披露主题鲜明集中，主要承担科学传播任务。

[Purpose/significance] The research aims to quantitatively measure the information disclosure level of mul?
tiple subjects in public health emergencies and reveal the status of information disclosure during epidemics in China
from the data level. It provided a new analytical framework for the research on the information disclosure behavior of
multiple subjects and provided theoretical support for the government and other related entities to improve information
services. [Method/process] The article took the microblogs related to the epidemic disclosed by each subject as the re?
search object. It constructed an analysis model for the evolution of information disclosure contents and characteristics
of multiple subjects. Then, we used the TFIDF-LDA model to extract topics and measured the breadth and depth of in?
formation disclosure based on the theor y of social penetration. Finally, we analyzed the characteristics of information
disclosure content based on the life cycle theory and explored the information disclosure model and evolution trend of
multiple subjects. [Result/conclusion] The information disclosure content of multiple subjects during the epidemic was
on a wide range of topics, and there were significant differences in the content and characteristics of different subjects'
disclosure. The government's disclosure was positive and of the highest quality. It aimed at positively guiding public
opinion. Mainstream media had a wide range and depth of information disclosure and were highly sensitive to data priva?
cy. Self-media had a distinctive and focused disclosure theme and mainly undertook scientific communication tasks.