具有先验信息约束的半参数回归模型及其虚拟观测值解法

SEMI-PARAMETRIC REGRESSION MODEL CONSTRAINED WITH PRIORI INFORMATION AND ITS VIRTUAL OBSERVATIONS SOLUTION

针对半参数回归模型建模过程中合理性不足问题,提出了具有先验信息约束的半参数回归模型,并给出了该模型相应的虚拟观测值解法和正则化矩阵与正则化参数在测量中的物理意义;证明了该新解法与补偿最小二乘法的等价性以及该新模型的有效性和合理性。

In order to improve the rationality during the modeling, we proposed the semi-parametric regression model constrained with priori information. At the same time, we deducted the virtual observation solution for this model, which made the regular matrix and normalized parameter have physical meanings in surveying and mapping, and we also proved that the new method is equal to compensated least squares and the new model has good validity and rationality.