A Study Of COVID-19 Data In China Based On Functional Data Analysis And Epidemic Model

In early 2020,An epidemic caused by 2019-n Cov broke out in China,and coincided with the Spring Festival.The epidemic of pneumonia was rapidly spreading across the country.COVID-19 has had a huge impact on China's economy.According to the National Bureau of statistics,China's GDP fell 6.8% in the first quarter of 2020.While remarkable achievements have been made in the prevention and control of the epidemic,the epidemic situation in the world continues.After April,there continue to be imported cases in China,and there is still the risk of imported cases in China.This paper uses functional data analysis method and epidemic model to analyze the data from two scales of time and space in COVID-19,China,and provides suggestions for the evaluation and prevention and control of China's epidemic prevention and control measures.Firstly,a robust functional principal component clustering analysis method based on MCD algorithm and a robust functional principal component clustering analysis method based on L?2,the functional clustering analysis method of the distance is used.The data of China's COVID-19 are analyzed.The results of the analysis of simulated data and China's COVID-19 epidemic data show that there are differences in the preventive measures between the mainland and Hong Kong,Macao and Taiwan.This is related to the political and cultural differences between the regions.The analysis of the whole data shows that the epidemic situation between regions is different.In general,the proposed method can get robust and more explanatory analysis results.Secondly,this paper improves the SIR model from two aspects of time-varying parameters and scale parameters,and proposes a dynamic parameter variable scale SIR model,in which the time-varying parameters can reflect the changes of parameters with time in the study period,and scale parameters can be added to make the model adapt to the infectious disease data under different growth scales,and explains the feasibility and effectiveness of the improvement from two aspects.First,in order to satisfy the model assumption that there is no population flow in the SIR model,we use the logistic model and the generalized logistic model with scale parameters to analyze the COVID-19 data in Wuhan city during the closed season.The results show that the scale parameters are effective in the epidemic model of Hubei.Second,the fitting results of SIR model show that the parameters are not fixed in the analysis period,and timevarying parameter model is needed.In general,this paper improves the SIR model of infectious diseases by combining functional data analysis method.It also analyzes the two dimensions of China's COVID-19 data from time and space,evaluates the effect of China's epidemic prevention and control in the past,and puts forward some suggestions for the prevention and control of epidemic situation in the later stage.At the same time,the method used in this paper can also be applied to the global epidemic data.However,the quality of COVID-19's data varies from country to country,and the population mobility among countries is large.How to improve the robustness of the method deserves further study.