The linear model and hypothesis. A general unifying theory (Q901376)

The book deals with the classical topic of multivariate linear models. Linear regression and related statistical models belong to standard curricula of many master studies programs and, therefore, are covered in numerous textbooks. Thus, the results collected in the book are not particularly new. On the other hand, the monograph (and the book is definitely a monograph) presents a unified consistent matrix related treatment of all topics, which positively distinguishes it from other books on the subject. Readers who have at least some working knowledge of the facts proved in the book can fully appreciate the intellectually precise beauty of this unified approach. Probably this was the scope of the monograph in the first place. A few words about the content. The book contains all important kinds of regression models including nonlinear regression. A large part of the book is devoted to linear regression, estimation, inferences and various hypothesis testing. Some attention is paid to nonlinear regression and multivariate models. A considerable part of the book is devoted to the Large Sample Theory. The final chapter deals with the multinomial distribution. One of the nice features of the book is that all chapters are supplied with a relevant reference list. For the readers who are not so well acquainted with linear models the reading might turn out to be quite challenging. First of all, the definitions and assumptions for the results are not sufficiently emphasized and are usually presented in a manner which is somewhat matter-of-fact, clearly assuming that the readers already know what they are dealing with. Since assumptions are essential for the proofs, the repetitive process of looking for them in the text becomes quite difficult. The proofs themselves are very concise, more like in the journal papers, so there is quite a good chance to get stuck in the middle of the proof more than once. All in all, the monograph is a consistent, logical and comprehensive treatment of the theory of linear models aimed at scientists who already have a good knowledge of the subject and are well trained in application of matrix algebra.