向量时序SVM-AR模型在公路软土地基沉降中的预测研究

为提高公路软土地基沉降预测的精度,提高工程的安全性,文中分别探讨了SVM模型和时序AR模型适用范围和特点,并结合2类模型各自优点提出SVM-AR模型。该模型用SVM模型预测趋势沉降量,用AR模型预测随机沉降量,然后组合获得预测沉降量。工程实例表明,SVM-AR比SVM模型预测结果更为准确,更好地反映公路软土地基沉降过程。

Study on the Prediction of Soft Highway Foundation with Vector Time Series SVM-AR Model

In order to improve the prediction accuracy of highway soft foundation settlement,and improve the safety of the engineering,the scope and characteristics of SVM model and time series AR model are discussed respectively,and a new method SVM-AR model is put forward by combining the respective advantages of the two models.In the new model,the trend settlement and random settlement are predicted by SVM model and AR model respectively,and the total settlement is combined by them.Engineering example shows that the SVM-AR prediction results are more accurate than that of the SVM model and better reflects the settlement process of highway soft soil foundation.