支持向量机的核函数及其参数选择对大坝监控模型预测精度的影响

Influence of SVM Kernel Function and Parameters Selection on Prediction Accuracy of Dam Monitoring Model

鉴于支持向量机的核函数及其参数选择对大坝变形监控模型的预测精度有重要影响,基于支持向量机常用的高斯核函数、多项式核函数及其线性混合核函数分别建立了大坝监控模型,并利用改进的粒子群算法对支持向量机的惩罚因子和核参数进行寻优。实例分析结果表明,基于高斯核函数的大坝监控模型的拟合能力较好,基于多项式核函数的监控模型的泛化能力较好,基于混合核函数的监控模型的综合性能较好。

The reasonable choice of kernel function and its parameters have an important impact on forecast accuracy of monitoring model for dam deformation.Based on Gauss kernel function,polynomial kernel function and linear mixture kernels function,the monitoring model of dam was established respectively.And then improved particle swarm optimization is used to optimize penalty factor and kernel parameter of the SVM model.The experimental results indicated that the dam monitoring model based Gaussian kernel function had better fitting ability,the dam monitoring model based polynomial kernel function had better generalization ability and the dam monitoring model based mix kernel function had better comprehensive performance.