Comparing distribution of helminths between populations using stool-based multiplex real-time PCR

Abstract

Background: Malaria is still a major public health problem in China, despite the fact that the government has implemented a series of strategies to fight against malaria. Advanced warning and reliable forecasting can help policymakers to adjust and implement strategies more effectively,which will lead to the control and eradication of malaria. We explore the application of a dynamic nonlinear autoregressive neural network (NARNN) model in forecastingthe incidence trend ofmalaria. Methods &Materials:Monthly incidence cases of malaria from January 1986 to June 2012 were obtained from the Jingmen Center for Disease Control and Prevention in Hubei, China. Data from 1986 to 2011 as a time series was used to develop a NARNNmodel. The error autocorrelation plot, the time series response plot, the mean square error (MSE) and the correlation coefficient (R) were analyzed to choose the optimalmodel. The one-step-ahead prediction and multi-step-ahead prediction were conducted to validate the chosen model using the data from 2012. The mean absolute percentage error (MAPE) was used to evaluate the prediction performance. The NARNN modeling was implemented using the Neural Network ToolboxTM in MATLAB Version 7.11 (R2010b). Results: The most appropriate network we found applied to forecast themonthly incidence cases ofmalaria had13hiddenunits and 6 delays. The MSE of training set, validation set and testing set were 150.05, 128.08, 260.18, respectively and the R was 0.98. The correlations except for the one at zero lag, all fell within the 95% confidence limits around zero. The outputswere distributed evenly on both sides of the response curve of the neural network. The MAPE of one-month-ahead, three-month-ahead and six-monthaheadwere 0.78, 1.81, 3.12, respectively. The curve of observations and predicted values from 1986 to 2012 showed a good fitting effect. Conclusion: The dynamic NARNN model seemed feasible to track the incidence trend of malaria, which provides a theoretical basis for predicting and detectingmalaria in the study area. It is worth attempting to utilize the method in other malaria-endemic areas and other infectious diseases.