局部均值分解与支持向量回归的大坝变形预测  

Forecasting model of dam deformation based on LMD-SVR method

作　　者：王奉伟[1,2] 周世健[3] 周清[1] 池其才 

机构地区：[1]东华理工大学测绘工程学院,南昌330013 [2]流域生态与地理环境监测国家测绘地理信息局重点实验室,南昌330013 [3]南昌航空大学,南昌330063

出　　处：《测绘科学》2016年第10期132-135,共4页Science of Surveying and Mapping

基　　金：国家自然科学基金项目(41374007);测绘地理信息江西省研究生创新教育基地项目(2310700008)

摘　　要：针对现有变形预测方法对于大坝变形的预测效果不理想的问题,该文利用局部均值分解方法获取生产函数分量并进行支持向量回归建模,用此方法对大坝变形进行多尺度分析。通过局部均值分解对大坝变形序列进行分解得到其乘积函数分量,然后利用支持向量机回归进行外推预测,再把各乘积函数分量的预测结果进行叠加重构生成,进而获得大坝变形预测值。通过实例分析,比较GM(1,1)、支持向量机和该文方法3种模型在变形监测数据处理中的拟合和预测结果,表明该文方法充分发掘数据本身所蕴含的物理机制和物理规律,提高了大坝变形多尺度预测精度。This paper analyzed the theory of LMD and support vector regression, then proposed the forecasting model of darn deformation named LMD-SVR. By the LMD method, the deformation sequence was decomposed and then each subsequence was forecasted with SVR. After reconstruction, the forecasted results were obtained. Experimental results of three models showed that the LMD-SVR model was better than other models; it was able to perform dam deformation in different feature scales simultaneously with those characteristics. So the LMD-SVR model is very suitable to be applied to the dam deformation analysis and prediction of the deformation object.

关 键 词：局部均值分解LMD 支持向量回归SVR GM(1 1) 支持向量机SVM 变形预测 

分 类 号：P208[天文地球—地图制图学与地理信息工程]