The comparison of ARMA, exponential smoothing and seasonal index model for predicting incidence of Newcastle disease

Objective: The mathematical model prediction has been widely recognized among many disease forecasting methods. But different models will show out distinct prediction results of different diseases, as well as their occurrence locations. In this study, ARMA, exponential smoothing and seasonal index model were adopted for predicting incidence of Newcastle disease, and evaluating the precision of these three models. A predictive model was confirmed for the occurrence of Newcastle disease in the province A of China, through comparing with the three mathematical models. Method: ARMA, exponential smoothing and seasonal index model were established and fitted according to the incidence of Newcastle disease in province A from Jan. 2000 to Dec. 2007. Meanwhile, used the three models to predict the incidence of Newcastle disease in province A from Jan. 2008 to Dec. 2008 and then compared their effect of fitting and prediction. The results showed that the predictive value of ARMA model was most close to the true number, and has the fittest predicting effect. The method of quantitative predictions was used in numerically simulating the incidence of Newcastle disease, which can reflect the transmitting tendency and incidence of Newcastle disease in recently years and provide theoretical basis for constituting optimum preventive strategies.