人工神经网络在上海市肾综合征出血热发病率预测中的应用

目的]应用人工神经网络的方法开展上海市肾综合征出血热发病率的预测。方法]采用广义回归神经网络和反向传播神经网络的方法,将上海市历史人群抗体阳性率、宿主动物的监测资料和气象数据作为训练样本进行上海市肾综合征出血热历史疫情拟合,并开展未来发病率的预测。结果]两种人工神经网络方法可综合监测资料,对上海市散发的肾综合征出血热的发病率进行拟合和预测,广义回归神经网络方法的拟合和预测效果优于反向传播神经网络方法。结论]人工神经网络方法可以用于上海市肾综合征出血热发病率的预测,上海市未来发病率可能保持在低水平。

Application of artificial neural network in forecasting incidence of hemorrhagic fever with renal syndrome in Shanghai

Objective] To explore the application of artificial neural network approach to forecasting the incident rate of hemorrhagic fever with renal syndrome（HFRS） in Shanghai. Methods] Approaches of generalized regression neural network（GRNN） and back propagation（BP） neural network were chosen in the study.The HFRS surveillance data on Shanghai population sero-positivity rate of HFRS antibody,and on density and infection rate of host animal plus meteorological data on Shanghai were treated as training samples,and epidemic trend of hemorrhagic fever with renal syndrome was forecasted. Results] Two artificial neural network methods integrated all kinds of surveillance data on hemorrhagic fever with renal syndrome in Shanghai with meteorological data on fitting and forecasting HFRS incidence.GRNN neural network in fitting and prediction was better than BP neural network. Conclusion] Artificial neural network methods are useful and effective in forecasting the incidence of HFRS in Shanghai,which may remain low in the future.