节点文献

使用思维进化算法优化的神经网络建立肾综合征出血热预测模型

Application of Mind Evolutionary Algorithm Optimized Neural Network Model to Predict the Incidence of Hemorrhagic Fever with Renal Syndrome

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 吴伟郭军巧安淑一任仰武夏玲姿周宝森

【Author】 Wu Wei;Guo Junqiao;An Shuyi;Department of Epidemiology,School of Public Health,China Medical University;

【机构】 中国医科大学公共卫生学院辽宁省疾病预防控制中心

【摘要】 目的探讨思维进化算法优化的BP神经网络在建立肾综合征出血热发病率预测模型中的应用前景。方法使用1984-2013年沈阳市的鼠情资料(鼠密度和鼠带毒率)和气象资料(平均气温、降水量和日照时数)作为网络的输入,同年的肾综合征出血热发病率作为网络的输出。把1984-2009年的数据作为训练样本,2010-2013年的数据作为预测样本。分别建立BP神经网络和思维进化算法优化的BP神经网络预测模型,并比较两种模型的拟合和预测效果。结果对于训练样本和预测样本,思维进化算法优化的BP神经网络的平均绝对误差(MAE)、平均绝对误差百分比(MAPE)以及均方误差平方根(RMSE)均小于未优化的BP神经网络。结论思维进化算法优化的BP神经网络预测模型的拟合和预测效果均优于未优化的BP神经网络,具有较强的推广应用价值。

【Abstract】 Objective To explore the application prospect of mind evolutionary algorithm optimized neural network model in building prediction model of hemorrhagic fever of renal syndrome.Methods Rat epidemic information including rodent density and viral carriage of rodents and meteorological data including average temperature precipitation and sunshine duration from 1984 to 2013 in Shenyang city were used as the input of neural network.The incidence of HFRS in the same year was used as the output of neural network.Data from 1984 to 2009 were selected as training sample,while data from 2010 to 2013 were selected as predicting sample.BP neural network and MEA optimized BP neural network were built respectively.Fitting and forecasting effect were compared between the two models.Results For the training sample and predicting sample,the mean absolute error,mean absolute percentage error and root mean square error of mind evolutionary algorithm optimized BP neural network were smaller than that of BP neural network.Conclusion MEA optimized BP neural network fitting and forecasting the HFRS incidence better than BP neural network,which is of great application value for the prevention and control of HFRS.

【基金】 国家自然科学基金项目(No.81202254,No.30771860)资助
  • 【文献出处】 中国卫生统计 ,Chinese Journal of Health Statistics , 编辑部邮箱 ,2016年01期
  • 【分类号】R512.8
  • 【被引频次】18
  • 【下载频次】319
节点文献中: 

本文链接的文献网络图示:

本文的引文网络