Epidemiological Characteristics Of Human Brucellosis In Shanxi Province And The Prediction Effect Based On ARIMA-ERNN Combination Model

Objectives:The epidemiological characteristics of human brucellosis in Shanxi Province from 2007 to 2017 w ere studied.The human distribution,temporal distribution and spatial distribution of human brucellosis in Shanxi Province were studied.Then,the ARIMA model,ARIMA-BPNN model and ARIMA-ERNN combination model are established,and the three models are used to predict.The modeling performance and predictionperformance of the model are evaluated to provide theoretical support for the prevention and control of human brucellosis in Shanxi Province.Methods:The monitoring data of human brucellosis in Shanxi Province from 2007 to 2017 were analyzed,and the three distribution characteristics were analyzed.The ARIMA-BPNN combination model,ARIMA-ERNN combination model and ARIMA model were established using the monthly incidence data of human disease in Shanxi Province from 2007 to 2017.And use the 2017 full-year data to evaluate the prediction effects of these three models.Results:1?Human distribution: The number of reported cases of human disease in 11 years is 3.75 times that of males,mainly distributed in the 40-65 age group.The patients are mainly farmers and herdsmen.2?Spatial distribution: 11 cities and 11 years report the largest number of cases in the order of Datong City,Zhangzhou City,Zhangzhou City and Jinzhong City;the top four reported incidence rates are Luzhou City,Datong City,Zhangzhou City and Jinzhong City.3?Time distribution: The reported incidence of human brucellosis in Shanxi Province in 2007-2014 is increasing year by year,while the trend is decreasing year by year in 2015-2017,the high incidence month from March to July,and the reported cases in May.The number is the most.4?The ARIMA-BPNN combination model,the ARIMA-ERNN combination model and the ARIMA model were established using the monthly incidence of human brucellosis in Shanxi Province from 2007 t o 2016,a nd the incidence rate in 2017 w as predicted.Among them,compared with the ARIMA model,the ARIMA-ERNN combined model has a MAE reduction of 18.65%;a MSE reduction of 31.48%;a MAPE reduction of 64.35%;a MAE reduction of 60.19% and a MSE reduction of 75.30%;MAPE Reduced by 64.35%.Compared with ARIMA-BPNN,the combined model ARIMA-ERNN reduced MAE by 9.60%;MSE decreased by 15.73%;MAPE decreased by 11.58%;MAE decreased by 31.61%;MSE decreased by 45.79%;MAPE decreased 29.59%.Compared with the ARIMA model,the ARIMA-BPNN combined model reduced MAE by 10.08%,MSE by 16.68%,MAPE by 12.53%,MAE by 41.78%,MSE by 54.44%,and MAPE by 49.37.%.Conclusions:In 2007-2017,the number of reported cases of human brucellosis in Shanxi Province was more than three times that of females,mainly distributed in the 40-65 age group,mainly for farmers and herdsmen.The case report of human brucellosis in Shanxi Province was first reported in 2007-2017.The trend of decline after rising,from March to July each year is the high incidence month and peak in May;2007-2017 Shanxi human disease patients mainly in Datong City,Zhangzhou City,Zhangzhou City and Jinzhong City,reported the incidence The top four rankings are Luzhou City,Datong City,Zhangzhou City and Jinzhong City.The modeling performance and prediction performance of the combined models ARIMA-ERNN and ARIMA-BPNN are better than the single ARIMA model,and the combined model ARIMA-ERNN is optimal.