Dynamic Linear Models with R: Statistical Analysis & Time Series Forecasting for Data Science | Use R! - Perfect for Researchers, Economists & Data Analysts

I did use this book for the last 4 years.I believe that the quality of this book must be appreciated in context. As one reviewer stated (and I fully support that) - this is a very good description of how to apply the dlm package with well-selected examples. Its strengths are: (1) very well written package - I did learn some new things about R while analyzing the code; (2) good examples illustrating finer points of dynamic modeling in multiple contexts; (3) clear though sometimes terse explanations of the overall field.Since this is the best book on application I have found (through the years) it definitely deserves 5 stars. That does not mean it is perfect for everybody.I disagree with the 3-star review [with the exception of 'The software package is itself very powerful end elegantly implemented' :-)].This book is not be-all-end-all and it does not attempt to be. The theoretical basis and numerous - really numerous - and well explained practical examples are contained in 680 pages of 'Bayesian Forecasting and Dynamic Models' by West and Harrison. More recent 'Time Series: Modeling, Computation, and Inference ' by Prado and West contains plenty of explanations using similar methods and gives a good update on theory. There are many other books, though I found 'Time Series Analysis by State Space Methods' by Durbin and Koopman (2001 version) rather dry and tough going for somebody without earlier experience in this area.I think that the learning curve is slightly sharper in state space than in more traditional ARIMA-based approach. Also, the more traditional approach has simply many more books published on various level, most of that introductory. Still, if you want to use better version of modeling that is a small price you have to pay.Possibly some will expect this book to be more like (excellent in that area) 'An Introduction to Analysis of Financial Data with R' by Ruey Tsay which can be 'consumed' without much external reading and there is plenty of R examples to illustrate most of the element of the traditional approach. However, note that Tsay's book has an _introduction_ in the title, thus different audience. Additional advantage is the limitation of the topic to the financial time series - while 'Dynamic Linear Model with R' are for multitude of application areas.Personally I wish the authors found time to create a second edition of this book with some updates to the methods etc. - though I do appreciate that the market for such books is small. For general state space and dynamic modeling field a book similar in approach to Tsay's 'An Introduction to Analysis of Financial Data with R' could result in wider use of that approach.