Adopting Topic Modelling Approaches to Analyse Post-Pandemic Changes in Public Risk Perception

新冠肺炎疫情给社会生活带来了前所未有的变化。政府努力将疾病的传播降至最低，强调行为干预，包括鼓励社会疏远等保护性措施。公众参与保护措施可能与个人感染病毒的感知风险有关。虽然大多数研究大流行期间风险认知的文献都集中在横断面分析上，但很少注意确定纵向的社会变化。这项研究建议建立一个全面的框架，以分析大流行后公众风险认知的变化，这可能会导致广泛的社会变化。它旨在利用无监督的机器学习方法，如主题建模，来分析与疫情有关的非结构化社交媒体数据集。我们相信，这个项目将丰富以前专注于计算分析公众风险感知的研究。

The COVID-19 pandemic has caused unprecedented changes to social life. Government efforts to minimise the spread of the disease have emphasised behavioural interventions, including encouraging protective measures such as social distancing. Public engagement in protective measures may be linked to individuals' perceived risk of contracting the virus. While most of the literature studying risk perception during the pandemic has focused on cross-sectional analysis, little attention has been paid to identifying longitudinal social changes. This research proposes to develop a holistic framework for analysing post-pandemic changes in public risk perception which may engender widespread social change. It aims to leverage unsupervised machine learning methodologies, such as topic modelling, to analyse unstructured social media datasets related to the pandemic. We believe this project will enrich previous research focused on a computational analysis of public risk perception.