基于提升小波和LS-SVM的大坝变形预测

Dam Deformation Prediction Based on Lifting Wavelet and Least Square Support Vector Machine

提出了一种基于提升小波和最小二乘支持向量机的大坝变形预测方法,通过提升小波分析提取大坝监测数据效应量,分别对各效应量使用最小二乘支持向量机模型进行训练预测,再将合成各分量的预测结果作为最终的变形预测结果。算例结果表明,该方法较符合实际情况,具有很高的预测精度和良好的泛化能力。

Based on lifting wavelet and least square support vector machine,the method of dam deformation prediction is presented.Firstly,the effect size of dam monitoring data is extracted by lifting wavelet analysis.Then,the amount of each effect size is trained with least square support vector machine.Finally,the predicted results of various components are synthesized to be used as the final prediction result of deformation.The calculation result shows that this method is more in accord with the actual situation with high prediction accuracy and good generalization ability.