Ke-Wei Wang, Yu Wu, Jin-Ping Li, Yu-Yu Jiang
Wuxi Medical College, Jiangnan University, Wuxi 214122, China.
Zhongguo Xue Xi Chong Bing Fang Zhi Za Zhi. 2016 Jul 12;28(6):630-634. doi: 10.16250/j.32.1374.2016089.
To explore the effect of the autoregressive integrated moving average model-nonlinear auto-regressive neural network (ARIMA-NARNN) model on predicting schistosomiasis infection rates of population.
The ARIMA model, NARNN model and ARIMA-NARNN model were established based on monthly schistosomiasis infection rates from January 2005 to February 2015 in Jiangsu Province, China. The fitting and prediction performances of the three models were compared.
Compared to the ARIMA model and NARNN model, the mean square error (MSE), mean absolute error (MAE) and mean absolute percentage error (MAPE) of the ARIMA-NARNN model were the least with the values of 0.011 1, 0.090 0 and 0.282 4, respectively.
The ARIMA-NARNN model could effectively fit and predict schistosomiasis infection rates of population, which might have a great application value for the prevention and control of schistosomiasis.
探讨自回归积分滑动平均模型-非线性自回归神经网络（ARIMA-NARNN）模型对人群血吸虫感染率的预测效果。
基于中国江苏省2005年1月至2015年2月的月度血吸虫感染率，建立ARIMA模型、NARNN模型和ARIMA-NARNN模型。比较这三种模型的拟合和预测性能。
与ARIMA模型和NARNN模型相比，ARIMA-NARNN模型的均方误差（MSE）、平均绝对误差（MAE）和平均绝对百分比误差（MAPE）最小，分别为0.011 1、0.090 0和0.282 4。
ARIMA-NARNN模型能够有效拟合和预测人群血吸虫感染率，对血吸虫病的预防控制可能具有较大的应用价值。
