Comparative Analysis of Prediction Models of Global COVID-19 Pandemic
Objective The prediction of the fatality rate of COVID-19 pandemic is of great significance for in-depth understanding of the severity of the new coronavirus,rational allocation of medical resources,and targeted epidemic prevention strategies.Methods This study divides the development of the epidemic into four periods based on the dominant strain of the new coronavirus variant.Six countries including the United States,India,Brazil,Mexico,Peru,China,and the global average case fatality rate were selected as study subjects.Six models including the Grey Model,Exponential Smoothing Model,ARIMA,SVM,Prophet and LSTM are used for fitting and forecasting,the advantages,disadvantages and applicability of each model are discussed,and the model with the best effect is selected to forecast the fatality rate in the world and key countries.Results Model comparison shows that various models have their own advantages and disadvantages.It is predicted that the growth rate of the cumulative number of confirmed cases and cumulative deaths in most countries has slowed down,and the development trend has gradually stabilized.Conclusion The study suggests that traditional time series model is suitable for the prediction of stable development trend and limited samples,and the machine learning model is more suitable for fluctuating data,which can be used for large sample predictions.Depending on the features of these models,application can be extended to other fields.
NETL
NSTL
万方数据
