基于KPCA-LSSVM的公路软基路堤沉降预测  

Highway Soft Base Embankment Settlement Prediction Based on KPCA-LSSVM

作　　者：成传亮[1] 王淑娟[1] 罗德超[1] 

机构地区：[1]重庆大学自动化学院,重庆400030

出　　处：《重庆理工大学学报（自然科学）》2010年第6期78-82,共5页Journal of Chongqing University of Technology：Natural Science

基　　金：教育部博士点基金资助项目(102063720090013)

摘　　要：针对公路软基路堤沉降发生过程中多变量、强耦合、强干扰、大滞后的复杂特性,利用KPCA先将非线性数据投影到高维空间使其映射呈线性关系,应用PCA提取出映射数据的线性特征信息,间接实现去除原始数据噪声,以降低样本的维数,然后再利用最小二乘支持向量机进行建模的方法对软基路堤沉降进行预测。仿真结果表明:与最小二乘支持向量机及主元分析—最小二乘支持向量机建模方法相比,该核主元分析与最小二乘支持向量机结合的方法能够更准确地预测路堤沉降,且满足精确性和适用性的要求。For the complex characteristics with multi-variable,strong coupling,strong interference and large time-delay in the process of highway soft base embankment settlement,this paper presents a new method which uses the Kernel Principal Component Analysis（KPCA） to map the nonlinear input space into linear feature high dimensional space.Principal Component Analysis（PCA） is used to extract the linear feature information and indirectly remove the noises of original data to reduce the dimensionality of samples,and then the Least Squares Support Vector Machine（LSSVM） is used to build model for the prediction of highway soft base embankment settlement.The simulation results show that in comparison with the LSSVM and PCA-LSSVM,the model of KPCA-LSSVM can accurately predict the embankment settlement and satisfy the needs of the accuracy and the applicability.

关 键 词：软基路堤 沉降 核主元分析 最小二乘支持向量机 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TU17[自动化与计算机技术—控制科学与工程]