针对拱坝坝肩抗力体是坝体的重要受力部位,应用结合逐步回归分析方法的BP神经网络构建监测模型,以提高BP神经网络的泛化能力和模型的预测准度和精度,利用C语言编程训练,成功完成了预测,且传统模型预测数据的残差平方和大于改进后的残差平方和。实例分析结果表明,该监测模型可行、有效,并具有通用性。 According to the dam abutment resistance body is an important part of dam body, the BP neural network combining stepwise regression analysis method is used to construct the monitoring model to improve the generalization ability of BP neural network and the prediction accuracy and accuracy of the model. By using C Language programming training, the successful completion of the prediction, and the traditional model predictive data residual square sum is greater than the sum of the residual sum of the residuals. The case study shows that this model is feasible, effective and universal.