Generalized Least Squares (Q4823183)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 2110661
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Generalized Least Squares |
scientific article; zbMATH DE number 2110661 |
Statements
Generalized Least Squares (English)
0 references
27 October 2004
0 references
multivariate normal distribution
0 references
Wishart distribution
0 references
elliptically symmetric distribution
0 references
group invariance
0 references
general linear regression model
0 references
generalized least squares
0 references
best linear unbiased
0 references
maximum likelihood
0 references
restricted generalized least squares
0 references
Gauss-Markov theorem
0 references
nonlinear
0 references
covariance
0 references
AR(1) error model
0 references
Anderson model
0 references
loss function
0 references
risk matrix
0 references
efficiency
0 references
equivalence
0 references
Zellner estimator
0 references
heteroscedastic model
0 references
serial correlation model
0 references
seemingly unrelated regression model
0 references
growth curve model
0 references
concentration inequality
0 references
uniform bound
0 references
upper bound
0 references
0 references
0 references
0 references
0.88408136
0 references
0.8824222
0 references
0.87679076
0 references
Regression analysis has been one of the most widely employed and most important statistical methods in applications. Among a number of branches of regression analysis, the method of generalized least squares estimation based on the well-known Gauss-Markov theory has been a principal subject, and is still playing an essential role in many theoretical and practical aspects of statistical inference in general linear regression models. This book provides an up-to-date, self-contained introduction to the unified theory of generalized least squares estimators (GLSEs). The book covers in depth the lower and upper bounds approach pioneered by the first author, which is a powerful and useful tool for generalized least squares estimation. The book also contains chapter-ending exercises that make it ideal for self-study or as a course text. It can be divided into three parts, corresponding, respectively, to Chapters 1 and 2, Chapters 3 to 6, and Chapters 7 to 9. The first part provides the basics for general linear regression models and GLSEs. In Chapter 1, multivariate normal distributions, Wishart distributions, elliptically symmetric distributions, group invariance and their properties are discussed. In Chapter 2, general linear regression models are given and the properties of the GLSEs including conditions for unbiasedness and finiteness of second moments are established. Two empirical examples are provided in the last two sections of this chapter. The second part, the main part of the book, is devoted to a detailed description of the lower and upper bounds approach and its applications to serial correlation models, heteroscedastic models, and seemingly unrelated regression models (SURs). In Chapter 3, generalized least squares predictors are defined and its fundamental properties are described; a nonlinear version of the Gauss-Markov theorem is established under fairly mild conditions on the distribution of the error term. In the end of this chapter, an application to GLSEs with iterated residuals is presented. In Chapters 4 and 5, efficiency of GLSE with applications to SURs, heteroscedastic models and serial correlation models are discussed and several types of effective upper bounds are derived for the risk matrix of a GLSE. Empirical examples are presented at the end of each chapter. In Chapter 6, uniform bounds for the normal approximation to the distributions of GLSP and GLSE are obtained. The last part provides further developments (including mathematical extensions) of the results in the second part. Chapter 7 is devoted to making a further extension of the Gauss-Markov theorem, which is a maximal extension in a sense and leads to a further generalization of the nonlinear Gauss-Markov theorem proved in Chapter 3. In the last two chapters, some complementary topics are discussed. These include concentration inequalities, efficiency under elliptical symmetry, degeneracy of the distribution of a GLSE, and estimation of growth curves. The asymptotic equivalence of the estimators of \(\theta\) in the AR(1) error model and the Anderson model are given in the appendix Each chapter contains exercises. This book is intended for first-year graduate students in statistics, mathematics, econometrics, biometrics and other related fields. It can also serve as a very good reference book. One of the key features of the book is a concise and mathematically rigorous description of the material. It is noted that in this book few exercises involve data and most of them ask for proofs. It would be very helpful if more data exercises would be provided, and an appendix on use of statistical software would be added.
0 references