Application value of ARIMA model in the prediction of syphilis in Bengbu city

ObjectiveTo explore the feasibility of applying the seasonal autoregressive moving average (ARIMA) model for predicting the incidence of syphilis in Bengbu city, and provide reference in the prevention and control of syphilis.

MethodsARIMA model was used to fit the monthly incidence of syphilis in Bengbu from 2008 to 2016 by SPSS21.0 software, and the optimal model was established according to the BIC criterion.The incidence rate of syphilis from January to June 2017 was predicted, which was compared with the actual incidence in order to test the prediction effect.

ResultsThe monthly incidence rate of syphilis was well fitted by ARIMA (0, 1, 1) (1, 0, 0) 12.The model statistic Ljung-Box Q=16.726, P>0.05, and the residual sequence was white noise.The predictive value of syphilis incidence from January to June 2017 in Bengbu was well consistent with the actual value, and the actual values were within the 95% confidence interval of the predicted value.

ConclusionsARIMA model can well simulate the incidence of syphilis in Bengbu, and provide some reference basis for prevention and control of syphilis.