Available:*
Library | Item Barcode | Call Number | Material Type | Item Category 1 | Status |
|---|---|---|---|---|---|
Searching... | 30000000115570 | QA278.2 G78 1990 | Open Access Book | Book | Searching... |
On Order
Summary
Summary
Ridge type estimators are carefully derived as special cases of Bayes, Mixed, and Minimax estimators. The similarities and differences in the derivations and mathematical forms of the estimators resulting from these three different points of view are compared and contrasted. The efficiencies of the estimators with respect to different mean square error criteria are evaluated. Annotation(c) 2003 Book News, Inc., Portland, OR (booknews.com)
Reviews 1
Choice Review
Gruber does an excellent job of comparing and contrasting the development and properties of ridge-type estimators that can be derived by both Bayesian and non-Bayesian (frequentist) methods. The author suggests that previous research has described the estimators from only one of the following points of view: a special case of the Bayes estimator, a special case of the mixed estimator, special cases of the minimax estimators, and the solution to the problem of finding the point on an ellipsoid that is closest to a point in the parameter space. However, in this book mathematical formulations of ridge-type estimators from the Bayesian and non-Bayesian point of view are discussed, as well as relationships between different kinds of prior information. There are very detailed chapters on the Kalman filter and experimental design models. Gruber also includes a brief historical survey of work done in this area. Highly recommended for undergraduate and graduate readers with knowledge of statistics and matrix theory. -D. J. Gougeon, University of Scranton
