[论文解读] Detecting Community Depression Dynamics Due to COVID-19 Pandemic in Australia

Jianlong Zhou, Hamad Zogan+4 more|arXiv (Cornell University)|Jul 5, 2020

Mental Health via Writing参考文献 40被引用 6

一句话总结

本研究提出一种多模态与基于TF-IDF的分类模型，利用2020年1月至5月期间的Twitter数据，检测澳大利亚新南威尔士州（NSW）社区抑郁动态。研究发现，新冠疫情爆发后抑郁水平显著上升，政府封锁措施和全球病例激增进一步加剧了抑郁情绪，且在地方治理区域（LGA）层面存在显著差异，尤其在应对严重公共卫生紧急事件时，而非本地病例数量。

ABSTRACT

The recent COVID-19 pandemic has caused unprecedented impact across the globe. We have also witnessed millions of people with increased mental health issues, such as depression, stress, worry, fear, disgust, sadness, and anxiety, which have become one of the major public health concerns during this severe health crisis. For instance, depression is one of the most common mental health issues according to the findings made by the World Health Organisation (WHO). Depression can cause serious emotional, behavioural and physical health problems with significant consequences, both personal and social costs included. This paper studies community depression dynamics due to COVID-19 pandemic through user-generated content on Twitter. A new approach based on multi-modal features from tweets and Term Frequency-Inverse Document Frequency (TF-IDF) is proposed to build depression classification models. Multi-modal features capture depression cues from emotion, topic and domain-specific perspectives. We study the problem using recently scraped tweets from Twitter users emanating from the state of New South Wales in Australia. Our novel classification model is capable of extracting depression polarities which may be affected by COVID-19 and related events during the COVID-19 period. The results found that people became more depressed after the outbreak of COVID-19. The measures implemented by the government such as the state lockdown also increased depression levels. Further analysis in the Local Government Area (LGA) level found that the community depression level was different across different LGAs. Such granular level analysis of depression dynamics not only can help authorities such as governmental departments to take corresponding actions more objectively in specific regions if necessary but also allows users to perceive the dynamics of depression over the time.

研究动机与目标

• 调查COVID-19大流行对澳大利亚新南威尔士州社区层面抑郁水平的影响。
• 开发一种新型分类模型，用于在大流行期间检测用户生成的Twitter内容中的抑郁极性。
• 在地方治理区域（LGA）层面分析抑郁动态的精细化特征，以评估心理健康反应的区域差异。
• 评估政府实施的限制措施和重大公共卫生事件对社区抑郁水平的影响。
• 为公共卫生机构提供数据驱动的洞察，以支持对大流行相关政策措施和事件的心理健康后果的评估。

提出的方法

• 本研究使用2020年1月1日至5月22日期间从澳大利亚新南威尔士州用户收集的Twitter数据。
• 提出一种新型抑郁分类模型，整合来自文本、情绪、主题和领域特定信号的多模态特征。
• 应用词频-逆文档频率（TF-IDF）方法，从推文中提取显著的语言学特征以用于分类。
• 该模型将推文分类为抑郁极性水平，从而实现对抑郁趋势的时间与空间分析。
• 在州级（NSW）和精细的LGA层级上分析社区抑郁动态，以检测区域差异。
• 采用统计和时间序列分析，将抑郁趋势与确诊病例数、封锁事件及重大公共卫生紧急事件相关联。

实验结果

研究问题

• RQ1 澳大利亚新南威尔士州的社区抑郁水平在COVID-19大流行爆发及发展过程中如何变化？
• RQ2 社区抑郁水平的变化在多大程度上与州级封锁等政府限制措施相关？
• RQ3 地方治理区域（LGA）层面的抑郁动态在多大程度上与本地确诊病例数，而非更广泛的区域或全国事件相关？
• RQ4 重大公共卫生紧急事件（如“鲁比公主号”邮轮疫情）在多大程度上影响了LGA层面的社区抑郁水平？
• RQ5 结合多模态特征与TF-IDF是否能有效检测并追踪公共卫生危机期间社交媒体内容中的抑郁极性？

主要发现

• 新南威尔士州的社区抑郁水平在COVID-19大流行爆发后显著上升，2020年3月初出现明显上升趋势。
• 实施全州封锁措施与社区抑郁水平的可测量上升相关，表明限制性措施对心理健康有负面影响。
• 抑郁水平更受全国或州级确诊病例数急剧增加的影响，而非单个LGA内的本地病例数量。
• 限制措施的放松也导致抑郁水平上升，可能源于对社区传播风险增加的焦虑加剧。
• 地方性公共卫生紧急事件（如特定LGA内爆发大规模疫情）即使在本地病例数较低时，也对当地抑郁水平产生显著影响。
• 该模型成功检测到LGA之间抑郁动态的区域差异，部分区域（如里亚德）在特定日期（如2020年3月10日和16日）出现急剧的局部抑郁峰值。