摘要
本文报告我们承担的国家自然科学基金新冠专项项目的研究进展,其主要结果已发表在Nature(Jia等人[1])期刊,同时也报告一些相关的扩展分析。首先,基于在疫情爆发前武汉输入到全国各地的人口数量,建立了时空基准风险模型,成功地预测和解释了新冠疫情在全国各地的时空分布特征。进一步,为了在新冠疫情初期对不同地区的疫情风险进行评估,我们发展了一个社区传播风险指数以及风险探测的一套工具,而传播风险是根据实际确诊人数偏离基准模型预测的显著程度来测量的。这一指数可用于疫情预警系统,来辨识和追踪那些具有高传播风险的地区。最后,我们用统计模型和机器学习的方法评估各种人口流动风险源的异质性,检验他们对新冠疫情传播的相对贡献大小,包括比较武汉居民与非武汉居民、不同年龄段、不同性别等对疫情传播的影响。
This paper reports the research outcomes of our National Natural Science Foundation of China COVID-19 special project.Our primary results were published in Nature(Jia et al.2020).We also report extensions of this work.First,we report a spatio-temporal benchmark hazard model based on population outflow from Wuhan to other prefectures in China's Mainland before the lockdown,which successfully predicts and explains the spatio-temporal distribution of COVID-19 across China.Second,we report a risk index for community transmission risk that can be used as a risk detection toolkit assessing epidemic risk for different areas during the early stages of the COVID-19 pandemic.In particular,we estimated transmission risk based on the deviation between confirmed cases and the prediction of the benchmark model.This index can serve as an epidemic warning system to identify and track which regions have a high level of community transmission risk.Finally,we present statistical models and machine learning approaches that can assess heterogeneities in risk sources of population outflows and test the relative contributions of different risk sources on the spread of COVID-19;for example,we estimate the relative contribution of Wuhan versus non-Wuhan residents,men versus women,and different age groups on COVID-19 transmission risk.
作者
贾建民
袁韵
贾轼
Jia Jianmin;Yuan Yun;Jia Jayson S(Shenzhen Finance Institute/Shenzhen Institute of Artificial Intelligence and Robotics for Society/School of Management and Economics,The Chinese University of Hong Kong,Shenzhen 518172;School of Economics and Management,Tsinghua University,Beijing 100084;Faculty of Business and Economics,The University of Hong Kong,Hong Kong SAR999077)
出处
《中国科学基金》
CSCD
北大核心
2020年第6期667-674,共8页
Bulletin of National Natural Science Foundation of China
基金
国家自然科学基金项目(72042009和72074072)
深圳市人工智能与机器人研究院(2020—NT001)的资助。
关键词
人口流动
新冠肺炎
疫情传播
时空风险模型
风险分析
population flow
COVID-19
epidemic transmission
spatio-temporal hazard model
risk analysis