Study The Prediction And Control Of Syphilis Model In Bayingolin Mongol Autonomous Prefecture Of Xinjiang

Objective:In order to explore the feasibility of the time series model,least squares support vector machine model and epidemic dynamics model in syphilis research and application in Bayingolin Mongol Autonomous Prefecture of Xinjiang,enriching the theory of the predictive control model of infectious diseases,and finding a prediction model with higher accuracy.By predicting early understand the incidence trend of syphilis in Xinjiang Bazhou area,reveal the law of syphilis epidemic development,find the key factors affecting the prevalence of syphilis,which laid the foundation for the further development of Xinjiang Bazhou area syphilis early warning,also provide strong theoretical and quantitative basis for the prevention and control strategy.Methods:According to the characteristics of syphilis,ARIMA prediction model was established by factor decomposition,stationarity test,model identification,parameter estimation,model diagnosis and optimization.Using the established model to fit the monthly incidence of syphilis in 2008²014 years,and to predict the incidence of syphilis for a short time.Secondly,the idea of data mining was introduced into the syphilis prediction analysis by the support vector machine technology,the least squares support vector machine was selected to select the appropriate kernel function and determine the corresponding parameters.It fitted the syphilis monthly incidence data for 2008²014 years,and predicted the incidence of syphilis from January 2015 to June.Thirdly,Based on the theory of infectious disease dynamics and the transmission mechanism of syphilis,the dynamic model of syphilis transmission was established.The incidence of syphilis using the data in recent years.The report,by using nonlinear least squares estimation of model parameters,calculated the spread of syphilis basic reproduction number,long-term trend forecasted future syphilis epidemic,and analyzed the sensitivity of the parameters of the model,found out the key factors affecting the prevalence of syphilis.Results:Firstly,according to the factor analysis of the reported syphilis onset data,the results showed that the onset of syphilis was cyclical.The stability test showed that the sequence was a non-stationary time series.Therefore,the ARIMA model of syphilis was established by first order difference of the original sequence.By model identification,ARIMA?0,1,1??1,0,1?12?1,0,1?₁₂2 was a relatively optimal model.The mean absolute error percentage?MAPE?of the fitting accuracy index and the root mean square percentage error?RMSPE?of the model were 16.39%and 29.24%,respectively.The predicted data showed that there was no significant change in the incidence of syphilis from January 2015 to June.Secondly,before the establishment of LS-SVM model,the first 2008-2014 years of disease syphilis rate data were normalized,choose radial basis function as the kernel function and the regularization parameter and kernel width using 5-fold cross-validation,and the minimum error of the gamma and sigmma value were 222,0.000102,respectively.The fitting process,the MAPE=8.89%,RMSPE=19.12%of the LS-SVM model,predicted that the incidence of syphilis remained high in the period from January 2015 to June.Thirdly,according to the natural transmission mechanism of syphilis and the degree of sexual activity of the population,the infectious disease dynamics model was constructed.The estimated value of the parameter was obtained by numerical simulation,and the basic regeneration number was R₀=1.06?95%CI:1.01-1.15?.The MAPE and RMSPE of the model were within 10%,and the model was used to predict the last 10 years,and the results showed that syphilis was on the rise.The sensitivity analysis showed that the infection rate of?maximum?|PRCC|=0.9661?,followed by a core group of partners to change the amount of c₁?|PRCC|=0.8794?and the cure rate of?|PRCC|=0.6719?,that the infection rate of beta,a core group of partner number c₁ and the cure rate is the key factor influencingof syphilis.Conclusion:It is feasible to analyze and predict the epidemic situation of syphilis in Bayingolin Mongol Autonomous Prefecture of Xinjiang by three methods:ARIMA model,LS-SVM model and infectious disease dynamic model.The results showed that the incidence of syphilis has increased,should increase the rate of condom use revealed to cut off the spread of syphilis,prevention and intervention of the core groups of the network to strengthen the second,strengthen the sexual partners of Syphilis Serological Test,and found that the syphilis infection sources and may infect sexual partners,early treatment.