摘要
将进化支持向量机方法用于边坡变形规律的研究,用遗传算法搜索支持向量机最优参数,避免了人为选择支持向量机参数的盲目性,提高了支持向量机的推广预测能力。利用这种方法预测边坡变形规律,并与监测到的历史数据进行对比,以便工程技术人员及时调整设计方案和施工,维护边坡的稳定性。工程实例表明,该方法具有预测精度高和实时性等特点,具有广阔的应用前景。
The Method of genetic-support vector machine was applied to forecast the deformation of slope. It is shown that this method can avoid the blindness which is common phenomenon in man-made choice of parameters of support vector machine. It enhances the learning efficiency of the support vector machine and the capability of forecasting. Through comparison between the forecasting and monitoring deformations, it can help the engineer to modify the plan and keep the stabilization of slope. The engineering case analysis indicates that there is an extensive prospect for the genetic-support vector machine.
出处
《岩土力学》
EI
CAS
CSCD
北大核心
2003年第4期631-633,643,共4页
Rock and Soil Mechanics
基金
国家重点基础研究发展规则 (2002CB412708)
中科院知识创新重要项目(KJCX2-SW-L1-3) 资助
