摘要
目的建立用于麻疹预测的BP神经网络模型,对麻疹的年发病率进行预测。方法确定预测模型的基本结构,以1996-2010年全国麻疹的年发病率数据为训练样本,以2011-2012年的发病率数据为检验样本,采用改进的BP神经网络算法训练模型。并对2013-2017年麻疹的发病率数据进行预测。结果建立了预测模型,该模型在仿真预测样本点的平均相对误差为1.908%,在检验样本处的预测误差为2.332%,在所有预测点的平均相对误差为1.973%。并获得了2013-2017年全国麻疹的年发病率预测数据。结论所建立的BP神经网络模型具有良好的预测精度,适合用来进行麻疹的发病率预测,具有应用价值。
[ Objective ] To establish a BP neural network model to predict the occurrence rate of the measles. [Methods] The basal frame of the prediction model was established , the incidence rate of measles from 1996 to 2010 were trained samples, the incidence rate of measles from 2011 to 2012 were tested samples, trained this model by using improved BP neural network arithmetic. The incidence rate of measles from 2013 to 2017 was predicted by this model. [ Results ] The three layer network model was used. The average relative error of simulated predicted samples was 1.908%, the predicted error of tested samples was 2.332%. The average relative error of all predicted data was 1.973%. And the predicted incidence rate of measles from 2013 to 2017 have been obtained. [ Conclusions ] This neural network model has high prediction accuracy, it is suitable to predict the incidence rate of the measles.
出处
《中国现代医学杂志》
CAS
CSCD
北大核心
2014年第31期52-55,共4页
China Journal of Modern Medicine
基金
河南省软科学研究重点项目(No:102400440002)
河南中医学院"科研苗圃工程"项目(No:MP2014-07)
关键词
麻疹
BP神经网络
模型
预测
measles
BP neural network
model
prediction
