Prediction Of The Number Of People Infected With COVID-19 In The US And Evaluation Of Epidemic Prevention Policies Based On The SEI_DI_UQHRD Model

New coronary pneumonia is a disease caused by COVID-19,which can be spread through breathing.Since the discovery of the disease in Wuhan,China in 2019,the new crown epidemic has spread rapidly across the country,and then spread to the entire world,causing huge damage to human society.The United States is more serious in today’s world.Therefore,this article hopes to make a quantitative forecast of the number of people infected by the epidemic to provide important scientific data to the government and medical organizations,assist public decision-making and medical allocation of resources,which is of great significance for finding effective measures to fight the epidemic and restore the development of all aspects of society.The article first introduces the classic infectious disease warehouse model SIR model,and gives the equations of the model and the method of solving the parameters through the TRR algorithm.An empirical analysis of the number of people infected in the United States based on the SIR model found that although the model fits the existing data well,the calculated infection rate β=0.0011 and the removal rateγ=0.0002,which are far beyond the range of the two[0.2,1.5],This is greatly inconsistent with the actual situation and clinical data,and the short-term prediction effect of the model is poor,and the long-term prediction of the number of infections is exponential,which is obviously inconsistent with reality.Since the traditional SIR model has only three variables(S(t),I(t),R(t)} and two parameters {β,λ),it is too rough to describe the spread of the new crown epidemic with many influencing factors.Therefore,this article adopts the latest SEIDIUQHRD model that can evaluate the epidemic prevention policy based on the characteristics of the new crown pneumonia.This model was proposed by Khondoker Nazmoon Nabi in October 2020.It innovatively uses eight variables{S(t),E(t),ID(t),IU(t),Q(t)H(t),R(t),D(t)},fifteen parameters{β,λ,r1,r2,η,κ,q,σ1,σ2,γ,ΦD,ΦU,ΦH,δU,δH),etc.,corresponding to the introduction of the incubation period,asymptomatic infections,quarantined patients and quarantined rates,etc.can explain the spread of the new crown epidemic in detail and fully.We use this model to conduct empirical analysis on the United States,determine the initial values and approximate ranges of variables and parameters based on clinical data and past research,and use the TRR algorithm to fit the national data of the United States,and get very good results.The average absolute percentage error MAPE is 1.50%,and the predicted MAPE is 3.27%,which is much smaller than the 4.15%and 6.93%of the SIR model.At the same time,the model predicts the peaks and troughs of the epidemic for a long time,which is more realistic than the SIR model.At the same time,this article has also compared the predictions of other scholars,and the gap is within an acceptable range.Finally,we came to the conclusion that in the short-term it is estimated that there will be 33.49 million cases of symptomatic patients across the United States on March 31,2021,and the long-term U.S.epidemic will reach its peak in December 2021,and the maximum number of cases will reach 43.43 million around December 23,2021.,The number of subsequent infections began to decline slowly,and the number of infections may not be cleared within two years.It is expected that the US epidemic will become normal in the future.At the same time,in order to evaluate the impact of the epidemic prevention policy on the epidemic,we need to evaluate the impact of the isolation rate q,which is an alternative indicator of the epidemic prevention policy,on the number of symptomatic patients in the SEIDIUQHRD model.Traditional single-parameter or local sensitivity analysis is difficult to analyze models with such complex and uncertain parameters and inputs.We innovatively use Global Sensitivity Analysis(GAS)to evaluate the impact of isolation rate q on the model.Because the model is non-linear but monotonic,The partial rank correlation coefficient(PRCC)is used to analyze the influence of each parameter.Sure enough,the conclusion is that the isolation rate q is the parameter that has the greatest negative impact on the number of patients among many parameters.This shows that if the quarantine rate in the United States continues to be low,that is,if the American people continue to travel and march on a large scale,and the government’s epidemic prevention policy is still loose and people’s travel is not restricted,then the outbreak of the epidemic in the United States may be difficult to prevent.