Textbook in PDF format

This edition is about 100 pages longer than the first edition. The main reason for this is our desire to make the treatment more comprehensive. For example, we provide a more extensive foundation of filtering and smoothing for which we present proofs for a Bayesian analysis and for linear unbiased estimation as well as for a classical analysis. Our treatments for the linear model are based on four lemmas from multivariate regression theory. We have completely rewritten the discussions on simulation smoothing methods and we have added sections on dynamic factor analysis and state smoothing algorithms, including the Whittle smoothing relations and the two filter formula for smoothing. For a selection of chapters we have added a final section with exercises