针对高速公路建设中广泛存在的软土路基最终沉降量预测这一迫切需要解决的技术难题,提出了改进的BP神经网络预测模型,并归纳出影响软土路基沉降量的主要因素。通过对模型的建立、训练和验证,以及与其他方法的对比表明,改进的BP神经网络模型在非线性建模方面具有泛化性强、计算精度高、操作简便的独特优势,具有广阔的工程应用前景。 In view of the urgent technical problems that need to be solved to predict the ultimate settlement of soft soil subgrade in expressway construction, an improved BP neural network prediction model is proposed and the main factors that affect the settlement of soft soil subgrade are summarized. Through the establishment of the model, training and verification, and comparison with other methods, it shows that the improved BP neural network model has the unique advantages of generalization, high precision and simple operation in nonlinear modeling, and has a wide range of engineering Application prospects.