基于支持向量机的防护工程损伤评估  

Damage assessment method of protective projects based on support vector machine

作　　者：戎全兵[1] 程恺[1] 张睿[1] 靳大尉[1] 

机构地区：[1]解放军理工大学指挥信息系统学院,南京210007

出　　处：《计算机应用研究》2014年第6期1748-1751,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(70971137)

摘　　要：在对防护工程损伤影响因素分析的基础上,建立了基于支持向量机的防护工程损伤评估模型,完成从评估指标到损伤指数的非线性映射。结合历年来我国发生滑坡地质灾害情况时坑道工事的实际数据对模型进行了验证。计算结果表明,与误差反向传播神经网络相比,基于支持向量机的防护工程损伤评估模型能够再现专家领域知识和直觉思维,更准确地评估防护工程损伤程度,是一种有效的评估方法。Based on analyzing influencing factors of protective projects,this paper established the damage assessment model of protective projects in order to nonlinear mapping from evaluation index to damage index based on support vector machine,and validated the model with the actual data of tunnel works when the landslides happened over the years.Calculation results show that compared with error back propagation neural networks （BPNN）,support vector machine model which can reproduce the expert domain knowledge and intuitive thinking and assess the damage of protective projects more accurately is an effective evaluation method.

关 键 词：支持向量机 损伤评估 滑坡 坑道工事 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]