论文部分内容阅读
采用似乎不相关模型(SUR),可有效地考虑滑坡多个监测点的之间的关联性,对多个监测点的累计位移数据同步处理,更精确地实现滑坡的变形预测。以白水河滑坡为例,取ZG93和ZG118两个监测点,建立2个监测点的累计位移回归方程,根据两步回归法计算得到2个方程的误差协方差矩阵。由于误差协方差矩阵为非对角矩阵,所以2个监测点的回归方程实际上是相互联系的,满足似乎不相关模型的条件,可以建立2个监测点联立的似乎不相关模型,实现对ZG93和ZG118两个监测点的同步变形预测。与传统的普通最小二乘法(OLS)比较,似乎不相关的估计参数比普通最小二乘估计更接近真实值,SUR模型平均相对误差均小于OLS模型平均相对误差,显示SUR模型的预测精度要高于OLS模型。
The seemingly unrelated model (SUR) can effectively consider the correlation between multiple landslide monitoring points, synchronize the accumulated displacement data of multiple monitoring points and predict the landslide deformation more accurately. Taking Baishuihe landslide as an example, two monitoring points, ZG93 and ZG118, were established, and the cumulative displacement regression equation of two monitoring points was established. According to the two-step regression method, the error covariance matrix of two equations was obtained. Since the error covariance matrix is an off-diagonal matrix, the regression equations of the two monitoring points are actually interrelated to meet the conditions of the seemingly unrelated model, and the seemingly unrelated model of the simultaneous establishment of the two monitoring points can be established, Synchronous Deformation Prediction of Two Monitoring Points ZG93 and ZG118. Compared with the traditional OLS, the uncorrelated estimated parameters are closer to the real values than the ordinary least squares ones. The average relative errors of the SUR models are less than the average relative errors of the OLS models, indicating that the prediction accuracy of the SUR model is high OLS model.