摘要
本文提出一种基于最小二乘支持向量机(LS-SVM)的边坡稳定性预测方法,采用线性函数,多项式函数和径向基函数三种核函数,进行机器学习,经过反复计算和对比分析,建立精度较高的边坡稳定安全系数预测模型。以实例数据作为学习样本和测试样本,对模型进行检验,结果表明LS-SVM建模预测速度快,其预测精度与GA-BP神经网络算法和改进支持向量机算法(-νSVR)基本相近,将其用于边坡稳定性预测是可行的。
A slope stability predicting method based on least square support vector machine （LS-SVM） is proposed in this paper, in which three kinds of kernel functions （linear function, polynominal function and radial basis function） are used for the machine learning. With continuous calculation and comparative analysis, a high precision predicting model of slope stability coefficient is established and tested by some slope samples. These results show that the calculation speed of LS-SVM model is high, and the precision is close to GA-BP neural net algorithm and v- SVR algorithm. The proposed model can be applied to slope stability prediction.
出处
《水力发电学报》
EI
CSCD
北大核心
2009年第2期66-71,共6页
Journal of Hydroelectric Engineering
基金
国家高技术研究发展计划(863计划)专项课题(2007AA11Z121)
关键词
岩土工程
边坡稳定
最小二乘支持向量机
核函数
geotechnical engineering
slope stability
least square support vector machine
kernel function
