| Infectious diseases have never disappeared in the process of human development and have been paid close attention to all over the world.Infectious diseases are easy to cause global public health problems and have a great impact on human society.This study focuses on the transmission trends,spatial and temporal characteristics,risk factors and risk modeling of human infection with H7N9 avian influenza in China and COVID-19 in the United States.The main contents of this study are as follows: firstly,epidemic data and risk factors related to the epidemic are collected,and epidemic data sets are established.Then statistical analysis and spatial analysis were used to analyze the spread trend and spatiotemporal characteristics of the epidemic.Then the linear and nonlinear correlation analysis between epidemic and risk factors was analyzed.Finally,based on the previous research,the epidemic risk simulation model is constructed and optimized by combining the machine learning model and geographic information system for visualization.The results of this study are as follows:（1）From 2013 to 2017,human infection with H7N9 avian influenza continued to occur in China,with a total of 1,490 confirmed cases.Among them,the epidemic in2017 was the most serious,about 3.57 times the number of cases in 2013.The epidemic is more likely to break out in winter and spring,and the number of cases in December,January,February,and March accounted for 74.23% of the total number of epidemic cases.The epidemic presents a significant spatial and temporal clustering.The highrisk areas of the epidemic are mainly concentrated in the Yangtze River Delta region and the Pearl River Delta region.（2）From March 2020 to March 2021,the cumulative number of confirmed cases of COVID-19 in the United States was approximately 30.3 million.The epidemic is more likely to appear in winter and spring,in which the number of confirmed cases in November,December,and January accounted for 55.88% of the total number of cases.The epidemic mainly affected the eastern and central regions of the United States and showed significant spatial and temporal aggregation.（3）The degree of nonlinear correlation between human infection with H7N9 avian influenza and risk factors such as meteorological factors,social factors,and environmental factors in China is more significant.Among them,the degree of nonlinear correlation between the epidemic and the average maximum temperature is the highest,with a degree of freedom of 8.267.（4）The non-linear correlation between the COVID-19 outbreak in the United States and meteorological factors,air pollutants,and social factors is more significant.Among them,the non-linear correlation between the epidemic situation and the number of driving requests in Apple’s mobile trend data is the highest,with a degree of freedom of 8.863.（5）The risk simulation model of human infection with H7N9 avian influenza in China based on the multi-layer perceptron model and multi-risk factors has well fitted the nonlinear correlation between the epidemic and risk factors.The simulation effects of different risk levels are: the AUC value of lower-risk is 0.795;the AUC value of low-risk is 0.672;the AUC value of medium-risk is 0.853;the AUC value of high-risk is 0.825.（6）A COVID-19 epidemic risk simulation model in the United States based on multi-risk factors and logistic regression model,multi-layer perceptron model,decision tree model,random forest model,and Light GBM model,and constructed through dimensionality reduction and weighting combination strategies.The model fits the nonlinear correlation between the epidemic and risk factors well.The simulation effects at different risk levels are: the AUC value of low-risk is 0.932;the AUC value of medium-risk is 0.789;the AUC value of high-risk is 0.815;the AUC value of higherhigh risk is 0.906.The epidemic risk simulation model constructed in this study has achieved good simulation results and combined with geographic information system technology,the simulated and real values of the epidemic risk simulation model are visualized,which can intuitively reflect the degree of epidemic risk on the map.In this way,it can help the public health department to prevent and control the epidemic and effectively reduce the impact of the epidemic on human society. |
PDF Full Text Request