基于最小二乘支持向量机回归的基坑变形预测  

Prediction foundation pit deformation based on least squar support vector machine regression

作　　者：徐洪钟[1] 杨磊[1] 

机构地区：[1]南京工业大学土木工程学院,江苏南京210009

出　　处：《南京工业大学学报（自然科学版）》2008年第2期51-53,58,共4页Journal of Nanjing Tech University(Natural Science Edition)

基　　金：江苏省普通高校自然科学研究计划资助项目(05KJB560040);江苏自然科学基金资助项目(BK2006565)

摘　　要：将最小二乘支持向量机回归用于基坑变形预测.根据基坑位移的实测时间序列资料,应用最小二乘支持向量机回归建立了基坑位移与时间的关系模型.研究结果表明,最小二乘支持向量机回归用于基坑变形预测,具有较高的预测精度.与通常采用的BP神经网络相比,该方法具有预测误差小、计算快速、所需数据少等优点.The least squares support vector machine （LSSVM） regression was proposed to predict the deformation of foundation pit. On the basis of measured foundation pit displacement data with real time serie, the model of foundation pit displacement with time was built by LSSVM regression. The results show that LSSVM regression is an effective method to achieve high accuracy for predicting the deformation of foundation pit. The proposed modeling approach is superior to classical back-propagation （BP） neural network in terms of computation speed, forecast errors and less data needed.

关 键 词：基坑 变形预测 最小二乘支持向量机 支持向量机 

分 类 号：TU443[建筑科学—岩土工程]