基于逐类组合支持向量机的边坡稳定性预测研究  

Study on Prediction of Slope Stability Based on Termwise-combination Support Vector Machine

作　　者：匡野[1] 

机构地区：[1]成都理工大学地质灾害防治与地质环境保护国家重点实验室,成都610059

出　　处：《路基工程》2013年第5期73-76,80,共5页Subgrade Engineering

摘　　要：首次采用逐类组合支持向量机(TCSVM)方法,用于解决边坡稳定性预测的问题。模型是先用支持向量分类机(SVC)对边坡状态进行判识,然后用支持向量回归机(SVR)建立边坡安全系数预测模型,再用建好的模型对未知边坡的稳定性进行判别和安全性系数预测。利用模型对71个边坡实例中的61个进行学习,10个进行检验。结果表明:TCSVM对边坡安全系数的预测结果均优于SVM和PCA-SVM。表明逐类组合支持向量机方法提高了安全性系数预测的准确率,对边坡稳定性研究具有积极意义。In this paper, termwise-combination support vector machine （TCSVM） was firstly used as the solution for prediction of slope stability. As regards the model, the identification of slope state was carried out by the support vector classification （ SVC）, then the model predicting the safety factor of the slope was established by the support vector regression （SVR）, at last, the judgment of slope stability and prediction of safety factor were implemented by the established model. By use of this model, 71 slope cases were sampled, including 61 ones for studying and the other 10 for verification. The results show that the prediction results obtained from TCSVM are superior to those from SVM and PCA-SVM, demonstrating that termwise-combination support vector machine may improve the accuracy of the prediction and plays a positive role for research of slope stability.

关 键 词：边坡稳定 逐类组合支持向量机 边坡状态判识 安全性系数预测 

分 类 号：U416.14[交通运输工程—道路与铁道工程]