测绘工程
測繪工程
측회공정
Engineering of Surveying and Mapping
2015年
12期
48-52
,共5页
卢辰龙%匡翠林%陈晓林%鲁鹤松
盧辰龍%劻翠林%陳曉林%魯鶴鬆
로신룡%광취림%진효림%로학송
大坝变形预报%多变量灰色模型%背景值优化%总体最小二乘
大壩變形預報%多變量灰色模型%揹景值優化%總體最小二乘
대패변형예보%다변량회색모형%배경치우화%총체최소이승
dam deformation forecast%multivariable grey model%background value optimization%total least square
传统多变量灰色模型MGM(1,n)的背景值误差会使得求解的灰色参数精度降低。总体最小二乘是一种可以同时顾及到观测误差与模型系数矩阵误差的数学方法。基于此,引入TLS对传统MGM (1,n)模型的灰色参数进行修正。通过对某大坝变形数据试算,验证表明,该方法能够有效地提高变形预报精度。
傳統多變量灰色模型MGM(1,n)的揹景值誤差會使得求解的灰色參數精度降低。總體最小二乘是一種可以同時顧及到觀測誤差與模型繫數矩陣誤差的數學方法。基于此,引入TLS對傳統MGM (1,n)模型的灰色參數進行脩正。通過對某大壩變形數據試算,驗證錶明,該方法能夠有效地提高變形預報精度。
전통다변량회색모형MGM(1,n)적배경치오차회사득구해적회색삼수정도강저。총체최소이승시일충가이동시고급도관측오차여모형계수구진오차적수학방법。기우차,인입TLS대전통MGM (1,n)모형적회색삼수진행수정。통과대모대패변형수거시산,험증표명,해방법능구유효지제고변형예보정도。
The background value error of traditional multivariable grey model MGM (1 ,n) will make the precision of grey parameters decline .The total least square is a kind of mathematical method which can simultaneously consider the observation error and model coefficient matrix error .Based on this ,the MGM (1 ,n) model grey parameters are modified by introducing the total least square (TLS ) method in this paper .Through the trial of a dam deformation data ,the results show the proposed method can effectively improve the prediction accuracy .