Evaluation Of The Fitting Effect Of Three Epidemic Dynamics Models On The Epidemic Situation Of COVID-19

Objective:In this study,three infectious disease dynamics models,SIR,SEIR,and SEQIR were constructed to simulate the new crown epidemic data in Chongqing from January 24 to March 15,2020,to evaluate the fitting effects of each model and select the best model.Optimized model to analyze the impact of changes in different parameters in the model on the development of the epidemic,and to provide a reference for formulating epidemic prevention and control measures.Methods:The research data is selected from the daily epidemic report of the Chongqing Municipal Health Commission from January 24 to March 15,2020.SIR,SEIR,and SEQIR models were used for modeling,and the least square method was used to optimize the models,and the optimal parameters of each model fitting and the minimum SSE were obtained.According to the fitting data,the result of each model evaluation index is obtained,the fitting effect of different models is analyzed,and the best model is selected.By changing the parameters of the optimal model,the effects of different epidemic prevention methods on epidemic control were analyzed.Results:1.From January 24 to March 15,2020,there were a total of 576 confirmed cases of COVID-19 in Chongqing,including 7 deaths and 570 cured cases.The cure rate of COVID-19 in Chongqing is 98.96%,and the mortality rate is 1.04%.2.The fitting result of the SIR model is S₀=1427,β=2.3889×10^^（-4）,γ=0.1267,SSE=15851.60,R₀=2.69;the fitting result of the SEIR model is S₀=2443,E₀=200,β=9.7792×10^^（-5）,ω=0.1772,γ=0.2270,SSE=8882.66,R0=2.40;SEQIR model fitting results are S₀=1793,E₀=231,Q₀=177,β=1.19×10^^（-4）,ω=0.2681,κ=0.2037,γ=0.2039,SSE=7843.20,R0=2.89.3.The mean square error（MSE）of the fitting result of the SIR model is 304.84,the root mean square error（RMSE）is 17.47,the average absolute error（MAE）is 14.47,and the average absolute percentage error（MAPE1）from January 24 to March 15 is 99.35%,to 3.The average absolute percentage error（MAPE2）on the 11th was 15.70%.The MSE of the fitting result of the SEIR model was 170.82,the RMSE was 13.07,the MAE was10.48,the MAPE1 was 82.22%,and the MAPE2 was 11.11%.The MSE of the fitting result of the SEQIR model It is 148.91,RMSE is 12.20,MAE is 10.03,MAPE1 is 73.39%,and MAPE2 is 10.52%.4.The fitting results of the SEQIR model show that the basic regeneration number R₀in Chongqing on January 24,2020 was 2.89,and the effective regeneration number decreased to less than 1 on February 3,and fell to less than 0.3 on February 11.It fell below 0.1 on the 25th.5.By reducing the infection coefficientβ,the number of infected people can be significantly reduced.By increasing the removal coefficient,the peak of confirmed cases can be reduced,and the pressure on the medical system can be reduced.By increasing the vaccination rate and increasing the protection rate,the population’s immune level can be improved,and significantly reduce the number of infected people.Conclusions:In this study,the SIR model,SEIR model,and SEQIR model were used to fit and analyze the data of confirmed cases of new coronary pneumonia in Chongqing.The results show that the SEQIR model has the best fitting effect.The basic reproduction number in Chongqing on January 24 was 2.89.On February 3,the effective reproduction number fell below 1 and continued to decrease,indicating that Chongqing’s epidemic control effect was very good.The development of the epidemic could be effectively controlled by reducing the value of parameterβand increasing the value of parameterγby strengthening prevention and control efforts,early detection and treatment of patients and other measures.By increasing the population vaccination rate and the protection rate of the vaccine,the number of infected persons could be significantly reduced and the epidemic intensity could be weakened.