Semi-parametric Regression in Dam Observation Data Processing

Stepwise regression is a popular method used to model dam observation data to describe deformation patterns. Due to the multicollinearity among observable variables, some variables such as reservoir water level and temperature are often excluded in the regressor. A displacement model using semi-parametric regression was presented on the example of an arch-type dam in China, and the horizontal deformation analysis for a single point located on the dam was performed. The stepwise regression was penalized and refined by using the non-parametric component of semi-parametric regression. It is found that the semi-parametric regression has better fitting effects than stepwise.