基于IBRF神经网络的泥石流预测研究

Prediction of debris flow based on IRBF neural network

介绍了泥石流爆发的各种影响因素和泥石流预测基本原理,为提高预测的准确性与快速性,将免疫骤类算法与径向基函数结合,建立了泥石流预测的IRBF神经网络模型,并与BP网络预测模型和常规RBF网络预测模型进行了对比模拟实验。实验表明IRBF神经网络具有更高的预测准确性和更短的训练时间,该方法用于泥石流预测有较高的应用价值。

The causes and the basic principles of prediction of the outbreak of debris flows have been introduced in this article.To improve the forecast＇s accuracy and rapidity,by means of combining the immune algorithm and radial basis function,a debris flow prediction model was established,which based on IRBF neural network.A simulation experiment was made to compare the BP network prediction model and the conventional RBF network prediction model.The simulation experiment shows that the neural network IRBF have a higher prediction accuracy and a shorter training time,and this method for debris flow prediction have a detterpractical application value.