应用BP人工神经网络模型预测急性心梗死亡率与气象因素的关系

Application of BP artificial neural network model to predict relationship between meteorological factors and mortality of acute myocardial infarction

目的建立气象因素与急性心梗的智能神经网络预测模型,探索BP神经网络预测模型在气象因素与急性心梗死亡率关系的应用,为哈尔滨地区急性心梗的预防控制措施提供科学依据。方法利用MATLAB7.0软件中的神经网络工具箱及2008年的气象数据建立急性心梗死亡率的反向传播网络(BP神经网络)预报模型。结果经过11次学习和训练,神经网络模型误差为0.00763,达到允许范围内。BP神经网络模型的拟合结果显示,脑出血死亡率MAE为0.18,预测准确度为82.53%。结论 BP人工神经网络具有适应性强,对数据要求不高,自学习能力等突出优点,操作简便且易于掌握和应用。BP人工神经网络模型可以作为哈尔滨市急性心梗死亡预测的一种新方法。

Objective To establish intelligent neural network prediction model of meteorological factors and acute myocardial infarction and explore prediction model of BP neural network in the application of the relationship between meteorological factors and mortality of acute myocardial infarction. To provide the scientific basis for prevention and contrnl of acute myocardial infarction in Harbin region. Methods Using MATLAB7.0 software neural network Toolbox and meteorological data of 2008 to establish forecast model of acute myocardial infarction mortality with back-propagation network （BP neural network）. Results After 11 times learning and training, neural network model error was 0.00763, within the allowed range. Fitting of the BP neural network model showed that cerebral hemorrhage mortality MAE was 0.18 per cent, forecasting accuracy was 82.53%. Conclusion BP artificial neural network has strong adaptability, data requirements are not high, prominent advantages such as /earning ability, easy to operate , easy to learn and apply. BP artificial neural network model can be used as a new method for predicting mortality of acute myocardial infarction in our country.