| 王 锋,孙 杨,刘 慧.基于ISVM—逐步回归组合的混凝土坝变形监控模型[J].江西水利科技,2018,(2):85-90 |
| 基于ISVM—逐步回归组合的混凝土坝变形监控模型 |
| Concrete dam deformation monitoring mixed model built on thecombination method of ISVM-stepwise regression method |
| |
| DOI： |
| 中文关键词: 混凝土坝 ISVM 逐步回归方法 监控模型 水压分量 预测 精度 |
| 英文关键词: Concrete dam Support vector machine Stepwise regression method Monitoring model |
| 基金项目: |
|
| 摘要点击次数: 606 |
| 全文下载次数: 491 |
| 中文摘要: |
| 针对混凝土坝监控模型中监测数据较少、维数和非线性程度较高以及逐步回归方法预报值波动大，部分测值失真致拟合效果不佳等问题，鉴于支持向量机(SVM)能有效解决小样本、高维数以及非线性问题的能力，为此，基于改进SVM理论(ISVM)对大坝监控模型中的水压分量进行模拟，并采用逐步回归方法将模拟后的水压分量、温度分量与时效分量进行结合并模拟，建立了基于ISVM—逐步回归组合的大坝监控模型，将拟合和预报结果与逐步回归—逐步回归、ISVM—ISVM及逐步回归—ISVM组合的预报结果进行对比.结果表明，基于ISVM—逐步回归组合的混凝土坝安全监控模型拟合精度高，预测效果好，将更适于大坝监控模型的建立. |
| 英文摘要: |
| In view of the concrete dam monitoring data in monitoring model is less，high dimension and nonlinear degree as well as the method of stepwise regression forecast value volatility，part of the measuring value loss can cause such problems as poor fitting effect，support vector machine(SVM)can effectively solve the problem of small sample，high dimension and nonlinear.So，support vector machine was used in the simulation of the concrete dam monitoring model of hydraulic component，and stepwise regression method is used to simulate the water pressure component with a combination and simulation of the temperature component and aging component to build the dam monitoring model based on the combination method of SVM-stepwise regression，then give a comparison and analysis between stepwise regression-stepwise regression model，SVM-SVM model and stepwise regression-SVM model，results show that the SVM-concrete dam safety monitoring model of combination of stepwise regression fitting precision is high，the forecast effect is good，this model will be more suitable for dam monitoring model. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
