人民黄河Issue(5)：99-101,3.DOI:10.3969/j.issn.1000-1379.2014.05.031
基于 SVM-ARIMA的大坝变形预测模型
SVM-ARIMA Forecasting Model for Dam Deformation
摘要
Abstract
The monitoring data of the dam deformation is a nonlinear and non-stationary time series. After the introduction of kernel function,sup-port vector machine can solve the nonlinear problem effectively,so support vector machine was used to predict the dam deformation. In order to im-prove the prediction accuracy,ARIMA model was used to further analysis the residual series,therefore,SVM-ARIMA combination model was es-tablished to predict the dam deformation. The time series of dam deformation was divided into two parts:the trend term and the error term,they were predicted by SVM and ARIMA models respectively,the sum of the two parts was the final predicted value. At last,there was a comparison be-tween the measured data and the predicted one combined with a project. The result indicates that the accuracy of the SVM-ARIMA model is high, completely meeting the project needs.关键词
大坝变形/预测模型/支持向量机/ARIMAKey words
dam deformation/forecasting model/support vector machine/ARIMA分类
建筑与水利引用本文复制引用
沈寿亮,刘天祥,宋锦焘,姜彦作,梁睿斌..基于 SVM-ARIMA的大坝变形预测模型[J].人民黄河,2014,(5):99-101,3.基金项目
国家自然科学基金资助项目（51179066，51139001）；新世纪优秀人才支持计划资助项目（NCET-10-0359）；江苏省杰出青年基金项目（BK2012036）；水利部公益性行业科研专项（201301061，201201038）；江苏高校优势学科建设工程资助项目（水利工程）（YS11001）。 （）
