Aiming at the prediction of coal gangue embankment settlement,based on the non-linear mapping ability and learning ability of BP neural network,a momentum BP neural network with variable learning rate was proposed to predict the coal gangue embankment settlement of Chang'an Expressway.The embankment settlement model was established by using the measured settlement data.The model overcame the shortcomings of slow convergence speed and long training time of momentum BP neural network.At the same time,the postreg function was used to test the results of network training.The research results show ...
