scientific article; zbMATH DE number 1010456

This book is intended primarily for the use of a fairly large number of statisticians, physicists, engineers, applied mathematicians, econometricians, and others who generally shy away from dealing with integrals involving real-valued scalar functions of matrix argument because the integrations often require the evaluations of Jacobians of matrix transformations. These are usually linear or nonlinear matrix transformations which may contain special types of matrices such as diagonal, upper or lower triangular, symmetric, skew symmetric, orthogonal, hermitian, skew hermitian, unitary, semiunitary, and so on. Computation of the Jacobians in such matrix transformations is considerably involved. The author develops some techniques systematically so that anyone with a little bit of mathematical maturity and exposure to multivariable calculus can easily follow through the steps and understand (and apply in their own fields) the various methods by which the Jacobians in complicated matrix transformations are evaluated. Numerical analysts, computer scientists, and people working on nonlinear least-squares problems often need to look into the existence, properties, and inversions of Jacobian matrices. Such existence conditions, the types of transformations leading to a given Jacobian, conditions for a certain transformation to be unique, particular types of mappings such as the Samuelson map, and so on, are not discussed in this book. Topics such as Jacobians on algebraic structures and dynamical systems, the Jacobian conjecture, maps with special types of Jacobian matrices such as sparse matrices, parallel maps on hypersurfaces, Jacobians in differential equations and algebraic geometry, the decomposition of a Jacobian matrix for applications to specific problems in the study of energy levels in physics, game theory, time-series analysis, and so on, are also not treated in this book. However, some sample references to such topics are given at the end of Section 1.2 of Chapter 1. This book concentrates on the topic of the evaluation of Jacobians in some specific linear as well as nonlinear matrix transformations, in the real and complex cases, which are widely applied in the statistical, physical, engineering, biological, and social sciences. Outlines of the proofs of the various results are also given in all the chapters. The subject-matter of this book is divided into six chapters. Chapters 1 and 2 deal with matrix transformations and the associated Jacobians in the real case. After giving some preliminaries on matrix derivatives, Chapter 1 concentrates on linear and some smooth nonlinear transformations. Chapter 2 covers more complicated nonlinear transformations, in the real case, involving diagonal, triangular, skew symmetric, orthogonal, and semiorthogonal matrices. Singular value decomposition, and transformations containing many matrices and submatrices, are also discussed in this chapter. Chapters 3 and 4 deal with matrix transformations in the complex case. These complex cases are essentially analogous to the results in the real case discussed in Chapters 1 and 2. Transformations involving hermitian, skew hermitian, unitary, and semiunitary matrices are also considered. Thus the materials in Chapters 3 and 4 are parallel to those dealt with in Chapters 1 and 2. Chapter 5 deals with real-valued scalar functions of matrix argument in the real case. After introducing such elementary functions as the exponential, Gamma, and Beta functions, hypergeometric functions are defined and discussed briefly. Generalized hypergeometric functions in the class of Meijer's \(G\)- and Fox's H\(-\)functions are also introduced. Various approaches available in the literature for defining hypergeometric functions of matrix argument, such as the ones through the generalized Laplace transform, zonal polynomials, and the \(M\)-transform, are illustrated here. Some results on fractional integrals and fractional derivatives are also given. Scalar functions of many matrix arguments in the families of Appell's functions, Humbert's functions, Kampé de Fériet's functions, and Lauricella's functions are introduced and many of their properties are established in this chapter. Some applications in statistics and physics are also indicated. The material in Chapter 6 runs parallel to that in Chapter 5, but deals with real-valued scalar functions of one or more matrices in the complex case. Thus the complex analogues of most of the material in Chapter 5 are to be found in Chapter 6. The subject-matter in this book is developed reasonably slowly with numerous worked-out examples. Some exercises are also given at the end of each section. It is hoped that the material in this well-written book will form a good reference source for statisticians, engineers, physicists, econometricians, applied mathematicians, and people working in many other areas. Most (if not all) of the material in this book can also be used for a one-semester graduate-level course on Jacobians and functions of matrix argument.

Recommendations

Cited in

(only showing first 100 items - show all)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4337018)