Ridge Type-generalized Inverse Estimator of Parameters in Gauss-Markov Model and Its Optimality
Gui Qing · Developments In Surveying and Mapping · 2001
In paper, starting from the LS estimation, a new biased estimator of unknown parameters in Gauss Markov model, called ridge type generalized inverse estimator (RGIE), is generated by combining the singular value decomposition expression of the design matrix and the ordinary ridge estimation method. Some optimality, such as admissibility, superiority over the LS estimator, relative efficiency, numerical stability, is discussed and some important conclusions are obtained. The computational results demonstrate that the RGI estimator can improve the LS estimator when the design matrix is ill conditioned.