Time Series Analysis by State Space Methods (Oxford Statistical Science Series) [Hardcover]

James Durbin (Author), Siem Jan Koopman (Author)

Book Description

ISBN-10: 0198523548 | ISBN-13: 978-0198523543 | Publication Date: August 2001 | Edition: 1

This excellent text provides a comprehensive treatment of the state space approach to time series analysis. The distinguishing feature of state space time series models is that observations are regarded as made up of distinct components such as trend, seasonal, regression elements and disturbence terms, each of which is modelled separately. The techniques that emerge from this approach are very flexible and are capable of handling a much wider range of problems than the main analytical system currently in use for time series analysis, the Box-Jenkins ARIMA system. The book provides an excellent source for the development of practical courses on time series analysis.

Editorial Reviews

Review

... provides an up-to-date exposition and comprehensive treatment of state space models in time series analysis. Journal of the Royal Statistical Society This book will be helpful to graduate students and applied statisticians working in the area of econometric modelling as well as researchers in the areas of engineering, medicine and biology where state space models are used. Journal of the Royal Statistical Society ... a good mixture of theory and practical applications ... graduate and research students will definitely enjoy this book. Also practitioners will find the book quite useful. I would also recommend it for library purchase. Journal of the Royal Statistical Society

About the Author

James Durbin is at London School of Economics and Political Science. Siem Jan Koopman is at Department of Econometrics, Free University, Amsterdam, The Netherlands.

Product Details

• Hardcover: 253 pages
• Publisher: Oxford Univ Pr (Sd); 1 edition (August 2001)
• Language: English
• ISBN-10: 0198523548
• ISBN-13: 978-0198523543
• Product Dimensions: 9.3 x 6.2 x 0.8 inches
• Shipping Weight: 1.1 pounds (View shipping rates and policies)
• Average Customer Review: 4.0 out of 5 stars  See all reviews (4 customer reviews)
• Amazon Best Sellers Rank: #975,741 in Books (See Top 100 in Books)

More About the Author

Discover books, learn about writers, read author blogs, and more.

Customer Reviews

Most Helpful Customer Reviews

2 of 2 people found the following review helpful:

3.0 out of 5 stars Not a particularly good book, September 23, 2010

By 

Amazonian - See all my reviews

Amazon Verified Purchase

This review is from: Time Series Analysis by State Space Methods (Oxford Statistical Science Series) (Hardcover)

Both authors of the book have authoritative stature in state space models. But this textbook is somehow stuck in a zombie land where it's neither fundamental enough to be an easy read like An Introduction to State Space Time Series Analysis (Practical Econometrics), nor in-depth enough to thoroughly cover more advanced topics such as non-Gaussian nonlinear state space models. Readers are simply directed to try Koopman's ssfpack (extended) or STAMP software, neither of which free.

5.0 out of 5 stars Good intro - not-so-good follow-up, June 14, 2011

By 

A reader - See all my reviews

This review is from: Time Series Analysis by State Space Methods (Oxford Statistical Science Series) (Hardcover)

Chapter 2 of this book must rank among the very best texts
ever written on the Kalman Filter: In a few pages, the authors
not only give a quick, comprehensable, implementable demo of
the Kalman filter (I had an implementation of the equations
up an working less than half an hour after I first opened the
book); they also motivate the various topics to be treated
in the rest of the book, like initialization, smoothing,
error control and so on.

Then... they fall through. While a lot of the simpler theory
is explained if not easily so at least comprehensable, the
authors tend to fall back on the 'we refer to the computational
package for further details' tretament way too often. Quite
frustrating to work through five pages of intense linear
algebra only to find that the crux of the chapter isn't
in there at all.

If there ever is a 2nd edition of this text, PLEASE make it
completely self-contained!

As for rating, the book as a whole might deserve 3 to 4 stars.
But that chapter 2... that chapter is worth 5 stars alone.

Easily.

4.0 out of 5 stars Good book, November 8, 2010

By 

Biostatistician - See all my reviews

Amazon Verified Purchase

This review is from: Time Series Analysis by State Space Methods (Oxford Statistical Science Series) (Hardcover)

State space models are a general and broad time series method and overcome the difficulty of dealing with the stationarity of the Box-Jenkins approach. All ARIMA models can also be stated and handled in state space models. In general, it is called Kalman filter in engineering and statistics.

Both authors are renowned researchers in time series analysis, especially in state space modeling. The book itself is mainly based on their publications and their colleagues' and is written from a statistical point of view. So many filters used in engineering such as extended Kalman filter (EKF) and sequential Monte Carlo (particle filter) were not included in it. There are two parts: Part I and Part II. Part I deals with linear Gaussian state space models including non-stationary time series analysis and one short chapter of Bayesian analysis. It's readable, but you should expect somewhat messy notations in some chapters. Part II deals with non-Gaussian and nonlinear state space models. Part II is solely based on both authors' seminal paper in 2000. Their paper in 2000 was cut significantly by the editor, so they took an opportunity to illustrate what was cut in detail in Part II. Bayesian analysis for non-Gaussian and nonlinear state space models is also included. Readers may have a little more difficulty reading Part II.

There are two main cons of the book. First of all, the coverage of non-Gaussian and nonlinear state space models is very limited because the treatment they introduced is just their paper in 2000. So readers cannot be exposed to other popular methods in engineering such as EKF and particle filter. Second, their computing tools are Koopman's software, which is commercial. So readers will find it hard to apply state space models for examples in the book.

However, in general, the book introduces the concept of Kalman filter nicely and rigorously.