论文部分内容阅读
为更准确预测煤与瓦斯突出强度,在组合算法和径向基函数(RBF)神经网络的基础上,建立变权重RBF组合模型。首先,选取最具代表性的3种单项模型:BP神经网络、支持向量回归机(SVR)、免疫遗传算法(IGA),分别建模后对样本序列进行预测,并重构预测结果数据。以重构后的预测序列为输入层,突出强度为输出层,对变权重RBF组合模型进行训练,获得各单项模型的动态权值,从而建立动态变权重RBF组合模型,最后对突出强度进行预测。结果表明:变权重RBF组合模型预测结果的平均相对误差为2.621 2%,优于各单项模型、定权重组合模型以及数据不重构组合模型。
In order to predict the intensity of coal and gas outburst more accurately, a combined variable RBF model is established based on the combined algorithm and radial basis function (RBF) neural network. First of all, we choose the most representative three kinds of single models: BP neural network, support vector regression (SVR) and immune genetic algorithm (IGA), predict the sample sequence after modeling, and reconstruct the prediction data. Taking the reconstructed prediction sequence as the input layer, highlighting the output intensity as the output layer, and training the variable weight RBF combination model to obtain the dynamic weights of each individual model, the dynamic variable weight RBF combination model is established, and finally the projected intensity is predicted . The results show that the average relative error of the prediction results of the variable weight RBF combined model is 2.621 2%, which is better than the single model, the fixed weight combined model and the data non-reconfigurable combined model.