论文部分内容阅读
目的探索径向基函数神经网络在甲型病毒性肝炎发病率预测中的应用价值。方法利用2004年1月-2009年12月重庆市法定报告的甲型病毒性肝炎月发病率资料分别构建径向基函数神经网络、BP神经网络模型,以2010年1-5月的发病率作为模型验证样本,对2010年6-12月的发病率进行预测,并对这两种模型的预测效果进行对比评价。结果径向基函数神经网络的预测结果的平均绝对误差(MAD)、平均相对误差绝对值(MAPE)和预测误差的方差(MSE)均小于BP神经网络。结论径向基函数神经网络的预测效果优于BP神经网络,对于甲型肝炎发病率预测来说不失为一条新颖而有效的途径。
Objective To explore the value of radial basis function neural network in predicting the incidence of hepatitis A virus. Methods According to the statutory monthly incidence of hepatitis A virus reported from January 2004 to December 2009 in Chongqing, a radial basis function neural network and BP neural network model were constructed respectively. According to the incidence rate of January-May 2010 as The model validation sample predicts the morbidity from June to December 2010 and compares the predictive effects of these two models. Results The mean absolute error (MAD), average absolute relative error (MAPE) and prediction error variance (MSE) of RBF neural network were all less than those of BP neural network. Conclusions Radial basis function neural network is better than BP neural network in predicting the incidence of hepatitis A, which is a new and effective approach.