基于加权多项式回归模型的变形预报

Deformation prediction based on weighted polynomial regression model

及时准确地预报变形是变形监测工作的核心内容之一,多项式回归模型以其考虑因素少,方程阶数低而得到广泛的应用。从预报角度来分析,距离变形预报时刻越近的观测数据对预报来说越重要。讨论了加权多项式回归模型的建立,重点介绍了权的确定,并结合生产实例验证了该法的有效性。

Timely and accurate prediction of deformation is one of the core contents of deformation monitoring,and polynomial regression model has been widely used because of few factors and low order of equation. From the perspective of forecasting analysis, the closer the observation data is to the deformation prediction time, the more important it is to the prediction. Therefore, this paper discusses the establishment of weighted polynomial regression model, focuses on the establishment of weights, and finally, the effectiveness of the method is verified by an instance.