小波神经网络在软基沉降组合预测中的应用

将小波神经网络组合预测模型引入软土路基沉降预测中。把5组不同形式的s型增长模型单项预测结果作为小波网络的输入向量,将代表相应时刻的实际值作为小波网络的输出,对软基沉降序列进行非线性组合预测。预测结果表明,小波网络组合预测的结果比各单项模型预测的结果都好,与BP神经网络相比,小波网络的收敛速度更快,预测精度更高,模型的泛化能力更强。

Application of Wavelet Neural Network Method to the Combination Forecasting for Settlement of Soft Ground

Using five types of S-type forecasting models as the input variables and using the actual values of the corresponding time as the output variables,the nonlinear combination forecasting model with the wavelet neural network is set up to predict the settlement of soft ground.The forecasting results show that the said model is obviously superior to other single models.They also show that the convergence speed of wavelet neural network is faster,the predicting results are more accurate and the generalization ability is better than that from BP network.