Public sentiment dynamics in policy transitions: a sentiment analysis based on Weibo data.

Introduction: China had been implementing stringent dynamic policies during the COVID-19 pandemic. In late 2022, China made a sudden policy shift from its three-year dynamic zero-COVID to the re-opening policy, which resulted in a divergence of online public opinions and varying sentiments. However, few research has been done to explore the public's sentiment changes toward this abrupt policy shift. Methods: To better inform effective health communication regarding governments' change of policies for future initiatives, this study aims to analyze public's sentiment changes toward the launching of China's re-opening policy by using Weibo data. Our study examined 1, 423, 694 Weibo posts during the period from November 11, 2022 to January 11, 2023 to conduct a fine-grained emotion extraction. This study also used the LDA topic model to extract potential topics in Weibo posts to align topics and corresponding emotions for generating in-depth understanding. Results: Fluctuations of different emotions during these two months were profoundly analyzed and interpreted by taking cultural, social, and policy-related reasons into consideration. Notably, the average proportion of "disgust" (24.0%) exceeded that of "like" (22.8%) after mid-December, while "happiness" exhibited a gradual increase to 12.0%. Discussion: Results of this study will be essential to informing the government's effective health communication in the time of public health crisis, facilitating pandemic control and prevention, and enlightening on the maintenance of public's well-being.