【摘要】 运用带适应学习率和动量因子的梯度递减法——TRAINGDX训练函数的BP网络对黄土的抗剪强度指标进行了预测。首先分析了影响黄土抗剪强度的6个影响因素,而后建立了6∶14∶2的神经网络(Artificial Neural Network)黄土抗剪强度指标的预测模型,最后借助Matlab为平台,利用自编的程序,进行了预测计算。通过对预测结果的分析可知该模型模拟和预测的精度均较高,可以应用到黄土抗剪强度指标的预测中。更多还原

【Abstract】 The shear strength index of Xi’an loess is predicted based on BP neural network with TRAINGDX training function in the gradient descent method.First,the six factors of shear strength are analyzed,then,the prediction model is determined with the structure of 6∶14∶2 for the simulation and prediction of the shear strength of Xi’an loess;finally,a computer program is made with Matlab language according to the improved BP algorithm.The results of prediction and computation show that the precision of simulation and prediction of this model is high,and it can be applied to the prediction for the shear strength index of Xi’an loess.更多还原