Research On Prediction Models For Two Types Of COVID-19 Infectious Diseases

The ongoing pandemic of COVID-19 had a serious impact on the world's economies,societies and the health of people.Under the strong leadership of the Communist Party of China,the Chinese government has successfully implemented the epidemic prevention and control system and measures,effectively curbing the spread of the epidemic.For other populated countries,epidemic prevention remains a huge challenge.It is of great research value and social significance to scientifically forecast the future development trend of the epidemic and make practical intervention decisions based on the available data of the epidemic.This paper first takes the epidemic data of provinces and cities released by the Health Commission of China as the research object,and proposed forward two types of prediction models to predicted and evaluated the development trend of COVID-19.(1)SE_qEIR(Susceptible-Isolated exposed-Exposed-Infected-Recovered)model for COVID-19epidemic prediction.The mutation of the virus leads to the concealment of the virus in highly infectious and asymptomatic persons,so two policy elements of intervention were considered:isolation and repeated nucleic acid testing.A more accurate prediction model SE_qEIR was proposed by improved the classical SEIR model.(2)COVID-19 epidemic prediction model based on LSTM.Second,the two models were used to forecast epidemic development in Kashgar,Xinjiang.(1)Kashgar used the improved SE_qEIR model to carry out prediction analysis in the studied area from four aspects:1)Simulated prediction of the number of infected people during the comprehensive nucleic acid testing of the population.2)Simulated prediction of the impact of isolation of asymptomatic patients on susceptible populations.3)Simulated prediction of the number of infected patients after increasing the isolation time and proportion due to the concealment of the virus.4)Control population movement and decrease the impact of person-to-person contact rate on the development trend of the epidemic.(2)LSTM model of deep learning was used to forecast the development trend of the number of infected people in Kashgar.Last,in order to verify the applicability of the improved SE_qEIR model and LSTM model,the two models were applied to another research area,Shanghai,to predict the trend of the number of infected people respectively.The error of the prediction results of the two models was analyzed.Experimental results show that compared with the LSTM model,the fitting effect and prediction performance of the improved SE_qEIR model were nearer to the real data under the condition of limited data.The SE_qEIR model can be used to reasonably estimate the trend of epidemic development.Due to the outbreak of domestic epidemic in local areas usually under the implementation of effective intervention policies will be quickly controlled until zero.Therefore,the improved SE_qEIR model can be used to analyze the trend of COVID-19transmission.At the present stage,there are still some problems in China's epidemic prevention policy,so we should combine the characteristics of the spread of mutant strains with them to formulate prevention,and control measures in the epidemic areas,and provide certain theoretical support for future epidemic intervention decisions.