×

Applied time series and Box-Jenkins models. (English) Zbl 0554.62073

New York-London etc.: Academic Press (Harcourt Brace Jovanovich Publishers). XII, 417 p. $ 25.00; £20.00 (1983).
The book is written for practitioners and for undergraduate students with very little statistical training. It is the author’s intention to enable these readers to understand the basic ideas of time series analysis and to enable them to construct and use ARIMA forecasting models.
The book is divided into 14 chapters and 3 appendices. The first six chapters deal with univariate time series analysis. After the introduction a brief discussion of stationarity is given. Chapter 3 presents ARIMA-models and the chapters 4 to 6 are devoted to identification (4), estimation and diagnostics (5), and forecasting (6), respectivley. Applications of time series analysis are given in the following three chapters 7 to 9. The first series to be analyzed is United States residential construction, followed by the United States unemployment series and the Clorox Company case.
After the three chapters of applied time series analysis the author now discusses bivariate (multivariate) time series approaches such as transfer function models and intervention models. Chapters 10 to 13 are concerned with transfer function models; after a brief introduction, identification of transfer functions (chapter 11) is discussed, followed by estimation and diagnostic checking (chapter 12) and by forecasting (chapter 13). The final chapter of the book presents intervention analysis.
For practical purposes the series analyzed in the book are reprinted in appendix A. Appendix B gives an overview of mathematical expectations, and in appendix C a primer of time series programs is given. References, a subject index as well as an author index are included.
The book is written in a simple language, and before presenting formulas the basic ideas of the formulas are discussed heuristically. Although the book is not sufficient for practical time series analysts it may nevertheless serve as a first introduction into the theory and practice of time series analysis.
Reviewer: H.S.Buscher

MSC:

62M10 Time series, auto-correlation, regression, etc. in statistics (GARCH)
62-01 Introductory exposition (textbooks, tutorial papers, etc.) pertaining to statistics
62M20 Inference from stochastic processes and prediction