Abstract:In traditional robust estimation, the selecting weight iteration method can not detect the systematic errors, while the semiparametric model based on the least square principle can separate the systematic errors well. This paper establishs the semiparametric regression model based the selecting weight iteration method, using the time series method and L-curve methods to determine the regularization matrix and smoothing factor, and uses the selecting weight iteration method to reshape the weight matrix so as to reduce the influence of gross errors and systematic errors in parameter estimation. Using a simulated example, and taking the measured data of Daping landslide in Fengjie county of Chongqing as an additional example, the validity and superiority of the selecting weight iteration for semiparametric regression method applied to landslide prediction in the Three Gorges reservoir area is verified.
YANG Yihui,ZOU Jingui,LI Qin et al. The Application of Selecting Weight Iteration for Semiparametric Regression Model in Landslide Prediction[J]. jgg, 2018, 38(8): 846-850.