Prediction of Incidence of Hand Foot Mouth Disease in Xinjiang Uygur Autonomous Region Based on ARIMA-GRNN Combined Model

Objectives Analyze the temporal distribution pattern of hand-foot-mouth disease in Xinjiang from 2009 to 2015 and predict of the incidence of hand-foot-mouth disease in Xinjiang in 2015 using autoregressive integrated moving average model-generalized regression neural network (ARIMA-GRNN), evaluate the effect of model prediction of incidence of hand foot mouth disease in Xinjiang.

Methods Using the Excel software to collect data on hand-foot-mouth disease in Xinjiang from 2009 to 2015, the hand-foot-mouth disease data from 2009 to 2014 were used as a training dataset, and R software was used to establish the ARIMA model. Using the Matlab 2014 software to fit ARIMA-GRNN combined model on the basis of ARIMA model. Using data in 2015 to test the effect of the fitted model.

Results As for ARIMA(0, 0, 1)(2, 1, 0)₁₂ model, the mean absolute percentage error (MAPE)=19.21, could match the incidence trend of hand-foot-mouth disease in Xinjiang well; as for the ARIMA-GRNN model built on ARIMA model, MAPE=15.625, could be better fitted, and the effect of the model was better than that of single ARIMA model. Then use the model to predict the incidence of hand foot mouth disease in Xinjiang in 2015, the predicted results are in line with the actual fluctuation trend of hand-foot-mouth disease.

Conclusions ARIMA-GRNN combined model could well fit and predict the incidence of hand-foot-mouth disease in Xinjiang. It has a positive effect on the prevention and control as well as monitoring of the disease.