Hi all. Today I recieved my copy of Durbin & Koopman: "Time Series Analysis by State Space Methods" Oxford University Press, 2001. http://www.amazon.com/gp/product/0198523548/sr=8-1/qid=1154184779/ref=sr_1_1/103-3640117-4140620?ie=UTF8 The book is aimed at an audience in statistics and econometrics, but serves as an intro to general Kalman filters. I like the general philosophy and structure of the book. As the authors state on page 1: "We ... decided ... to devote Chapter 2 of the book to a particularly simple example of a state space model ... and to develop as many as possible of the state space techniques for this model. Our hope is that this will enable readers new to the techniques to gain insights into the ideas behind state space methodology that will help them when working through the greater complexities of the general case." And do the authors succeed! Said chapter 2 starts on page 9, and withing page 13 -- in four pages -- have they not only sketched briefly the principles behind the Kalman filter, but also given the recursive equations for its implementation. The reader can download the example data from the given website, and try it himself. As a novice, it is possible to have a working Kalman filter within ten minutes from opening this book. Granted, some ideas are "dodgy" and some numbers in the initialization appear as out of thin air, but what a motivator for further study! After this initial example, everything else in this book is "merely" a matter of extending the KF to more dimensions, find efficient implementations and handle the realistic cases of initialization and estimation of the required parameters. I really like this style of presentation, so even if I as of right now have read only 16(!) pages, I have no reservations whatsoever against recommending this book as a primer in Kalman filters. Rune
Book review: "Time Series Analysis by State Space Methods"
Started by ●July 29, 2006