A hybrid model for short-term bacillary dysentery prediction in Yichang City, China

Bacillary dysentery is still a common and serious public health problem in China. This paper is aimed at developing and evaluating an innovative hybrid model, which combines the seasonal autoregressive integrated moving average (SARIMA) and the generalized regression neural network (GRNN) models, for bacillary dysentery forecasting. Data of monthly bacillary dysentery incidence in Yichang City from 2000-2007 was obtained from Yichang Disease Control and Prevention Center. The SARIMA and SARIMA-GRNN model were developed and validated by dividing the data file into two data sets: data from the past 5 years was used to construct the models, and data from January to June of the 6th year was used to validate them. Simulation and forecasting performance was evaluated and compared between the two models. The hybrid SARIMA-GRNN model was found to outperform the SARIMA model with the lower mean square error, mean absolute error, and mean absolute percentage error in simulation and prediction results. Developing and applying the SARIMA-GRNN hybrid model is an effective decision supportive method for producing reliable forecasts of bacillary dysentery for the study area.