中国全科医学Issue(8)：978-980,3.DOI:10.3969/J.iSSn.1007-9572.2015.08.031
手足口病短期预测的逆传播神经网络模型
Short-term Prediction of HFMD Based on Back Propagation Neural Network
摘要
Abstract
Objective To establish a back propagation neural network( BPNN)model for short-term prediction of HFMD to provide theoretical reference for its preventive strategy and measures. Methods A BPNN model was established. The monthly number of HFMD cases from January 2011 to January 2014 was taken as training model,that from February to March 2014 as test model,and that from April to september 2014 was predicted. Results The trained model was used to simulate and forecast the monthly numbers of HFMD cases from January 2012 to January 2014,showing that the average relative error absolute value was 0. 640 6%. The monthly number from February to March 2014 was used to test the model,the predicted monthly numbers were 39 783 cases,151 407cases,respectively,with an average relative error absolute value of 1. 814 7%. The average relative error of all predicting points was 0. 727 6%. BPNN model was used to predict the monthly numbers of HFMD cases from April to september 2014. The obtained predicted values were 298 567,454 730,412 590,270 988,126 758, 249 991,respectively,after normalization processing. Conclusion The variation trend of BPNN predicting monthly numbers of HFMD from April to september 2014 is identical with that over the same period of 2011—2013. The established model is of good predictive accuracy,which can be used to perform short-term prediction of HFMD.关键词
手足口病/神经网络/预测Key words
Hand,foot and mouth disease/Neural networks/Forecasting分类
医药卫生引用本文复制引用
徐学琴,孙春阳,刘晓蕙,裴兰英,王瑾瑾,栗彦芳,申杰,闫国立..手足口病短期预测的逆传播神经网络模型[J].中国全科医学,2015,(8):978-980,3.基金项目
河南中医学院“科研苗圃工程项目” （）
