Bayesian forecasting and dynamic models. (Q676677)

The review of the first edition from 1989, Zbl 0697.62029, still gives the main points. The new edition contains some new paragraphs: retrospective time series analysis, model estimation and diagnostics, the theory of limiting results in time series dynamic linear models (DLMs), new methods of time series decomposition in the state-space framework, time varying parameter autoregressive DLMs, and inference and application of autoregressive component DLMs. A new chapter focusses on the new and rapidly growing area of Markov Chain Monte Carlo approaches in dynamic models. Other parts are revised to incorporate new material and to make the text even more readable. Since the first edition, a companion volume by the authors together with \textit{A. Pole} has appeared [Applied Bayesian forecasting and time series analysis. (1994; Zbl 0855.62077)]. This is essentially a manual to the software package BATS, the implementation of the methods for the standard univariate Gaussian case described in chapters 10 and 11 of this book.