Correlation theory of stationary and related random functions. Volume I: Basic results (Q1262063)

This book presents another introduction to the spectral analysis of stationary stochastic processes. The term ``correlation theory'' in the title has to be understood as a synonym for ``second order theory'', i.e. restriction to first and second moments; the suggested considerable coverage of time domain methods is missing. Mixed autoregressive-moving average models are essentially used as a source for a parametric class of spectral densities only, whereas the Box-Jenkins methodology and state- space models are not mentioned at all. Chapter 1 provides the essentials on stationary stochastic processes and their covariance functions. Chapter 2 develops the spectral theory of such processes in discrete and continuous time, including a detailed discussion of linear filters. Chapter 3 discusses estimation of means, autocovariances and spectral densities. Practical aspects of spectral estimation are considered, but, often, the reader is referred to other literature, e.g., if he wants some explicit guidelines for choosing the smoothing parameter of nonparametric spectral estimates or for calculating autoregressive spectral estimates. So far, the content of this book is fairly standard. The last and longest chapter, however, reviews several related topics like random fields, generalized stationary processes and various spectral representations which are rarely mentioned in texts on time series analysis. The book aims at scientists and engineers who are interested in the applied aspects of stationary processes and who like an understandable, but also rigorous formulation of the underlying mathematical theory. Another part of the readership would consist of mathematicians and statisticians who want to learn something about the engineering applications of the spectral analysis of time series. The author succeeds in serving both audiences by stating clearly the main classical results, mostly with sketchy proofs or without any proofs at all, and by repeatedly describing and referring to particular applications. As the main results are not formulated as theorems, the text is a bit unstructed and comparably hard to read which is also due to the unsatisfactory typographical style. For a bridge from mathematical statistics to engineering, the book lacks the detailed bibliography which has been postponed to volume 2.