大坝变形预测的支持向量机模型  

Support Vector Machine Model of Dam Deformation Prediction

作　　者：张胜伟[1] 宋振柏[2] 张华荣[1] 李帅[1] 

机构地区：[1]山东理工大学建筑工程学院 [2]山东理工大学资源与环境学院

出　　处：《现代矿业》2012年第2期25-27,37,共4页Modern Mining

基　　金：山东省自然科学基金项目(编号:ZR2010DL002)

摘　　要：支持向量机在解决非线性及高维模式识别问题中表现出其特有的优势。针对大坝变形具有强非线性的特点以及传统神经网络模型预测时存在局部极小与过学习等问题,将支持向量机应用于大坝变形预测。以某大坝的变形监测数据为例,建立SVM预测模型,将支持向量机模型与神经网络模型进行比较分析。结果表明,该模型具有较高的精度,可以很好地应用于大坝变形监测分析。Support vector machine showed its specific advantage on nonlinear and high dimensional model recognition problem solving. Aiming at dam deformation had characteristic of strong nonlinear and problems like local minimum and over fitting exist in traditional neural network model prediction, support vector machine was used in dam deformation prediction. Took the deformation monitoring data of a dam as the example, SVM prediction model established, support vector machine model and neural network model were compared and analyzed. The result indicated, the precision of this model was high, which could be used in dam deformation monitoring analysis very well.

关 键 词：支持向量机 大坝变形 预测 神经网络模型 

分 类 号：TV698.1[水利工程—水利水电工程]