Regional Comprehensive Risk Assessment And Early Warning Of COVID-19

The COVID-19 outbreak in 2020 has affected more than 200 countries and regions around the world.As of 8 o ’clock on March 22,2021,Beijing time,the cumulative confirmed cases have exceeded 123 million and the cumulative deaths have exceeded 2.7 million,bringing a significant impact on the global economy.Now the epidemic is still spreading and has not been effectively controlled,some parts of the country are experiencing repeated outbreaks.Although the death rate of COVID-19 is low,it’s highly contagious.So it’s necessary to strengthen normal surveillance.Regional risk assessment and early warning during the spread of the epidemic can provide a scientific basis for the implementation of tiered prevention and control measures and the reduction of economic and social impacts.The existing regional risk assessment methods and research results are mostly from a single perspective,while the comprehensive perspective of research is less.Therefore,based on the risk assessment of public health emergencies and the theory of capacity and vulnerability,this paper constructs a comprehensive risk assessment and early warning model of COVID-19 from many aspects,such as the number of patients,population mobility,population structure,policy control,medical level and social economy.The research process of this paper is as follows:firstly,the risk assessment theories,empirical studies and prediction methods of COVID-19 and other infectious diseases at home and abroad are sorted out and summarized to lay a theoretical foundation for the research;then from the four dimensions of risk,vulnerability,exposure and disaster resistance,the comprehensive risk assessment index system and early warning model of regional COVID-19 were constructed,and the empirical analysis was carried out on 6 large cities in China and 21 prefecture-level cities in Sichuan Province from January 10,2020 to March 15,2020.Finally,countermeasures and suggestions for COVID-19 prevention and control were put forward based on the research conclusions.The main conclusions of this paper are as follows:（1）from the hazard,vulnerability,exposure and resistance against natural disasters four dimensions we select 13 indicators of risk,build new crown epidemic region comprehensive risk evaluation index system,and using principal component analysis dimension reduction,the main factors of comprehensive risk outbreak of the new champions league is: exposure degree and level of social development principal component factor,medical and health level and population outflow principal component factor,virus risk principal component factor and government control capacity and population flow principal component factors.（2）the empirical study,in view of Beijing,Shanghai,guangzhou,shenzhen,tianjin,chongqing’s new crown outbreak comprehensive risk index calculated result showed similar movement,highest comprehensive risk index after January30 th,after confirmed with patients in the hospital for treatment,as well as the government control in time,around the comprehensive risk index decreased gradually,February 14 after back to before the outbreak,after the country began to return to work in stages,to reopen,open public space,no longer limit public gathering,a slightly increased risk.Cluster analysis was conducted on the COVID-19 composite risk index of six cities,and it was found that the risk level of Beijing,Shanghai,Chongqing and Shenzhen fluctuated from low risk to high risk to medium risk.The change of risk grade in Tianjin is low risk – high risk –low risk – medium risk;The change of risk level in Guangzhou is low risk – high risk – medium risk – high risk.（3）In prefecture-level cities of Sichuan Province,the comprehensive risk level of COVID-19 in Ganzi Tibetan Autonomous Prefecture and Aba Tibetan and Qiang Autonomous Prefecture continued to be high risk during the 9 weeks from January 13,2020 to March 15,2020;The comprehensive risk of COVID-19 in Chengdu was low in the first two weeks of the epidemic,high in the third and fourth weeks,low in the fifth and sixth weeks,and medium in the last three weeks.The comprehensive risk level of the rest prefecture-level cities was low risk at the beginning of the COVID-19 outbreak,then reduced to low risk and remained so.（4）Taking Chengdu as an example,this paper adopts the GBDT model,Lightg BM model and XGBoost model in machine learning to conduct short-term prediction research on the constructed COVID-19 comprehensive risk assessment.It is found that GBDT model has the best short-term prediction effect,followed by Lightg BM model,and XGBoost model has poor prediction effect.