<< 1 >>
Rating: 
Summary: Most notable shortcoming is price
Review: I completely agree with the review 'A well written book for BJ modelers' (so please read it first). This book has some notable shortcomings, however. While very well done, it contents might be considered an edited paraphase of Box & Jenkins classic text entitled Time Series Analysis, and therefore adds very little to existing BJ literature. It often discusses solutions in terms of what "computer programs" will usually do for the model builder, but it does not reference specific programs or completely elaborate on how to create these programs from scratch. There are software packages out there that can help in this regard (MATLAB for instance), but this title was written in 1983 and predates all the modern model prototyping tools. It shows its age with graphs and tables generated from line printer output. The first 300 pages are "concepts", the last 300 are "cases" (as the title suggests), and there is significant overlap. This kind of repetition can be good if one is learning something for the first time, but it can also be deceiving if one were to assume the 600 page content to be as thorough as similarly-sized texts.It heavily references B&J 2nd edition with page numbers (sometimes quotes), B&J now being out in a later edition. As a result, the person that pays $200+ for this book (cost at the time of this review) will likely still need a copy of Box & Jenkins Time Series Analysis (and ultimately some kind of software). If you've already invested in a copy of B&J and found it unbearably intimidating or otherwise unintelligible, then this might be a book you can use. I would give it a strong 4 stars in that case, but I suspect that there may be less expensive books that might serve the same purpose. If the price ever comes down to under $50, this would be an excellent supplement to B&J.
Rating: 
Summary: A well written book for BJ modelers
Review: I found this book to be very well written. Everything is clearly explained and it covers just what the title suggests. Each chapter ends with a summary and there is one short chapter which contains a list of practical rules for setting up the forecasting models. Then there are 15 detailed case studies to show you how to put the theory into practice. Many plots are included showing the result of simulations. This is a BJ model builders book. If you are seriously interested in time series models of the unvariate type, consider this work. The book is geared toward a student or individual with a modest background in statistics.
Rating: 
Summary: nice text with many practical examples
Review: Pankratz gives a nice and non-rigorous description of Box-Jenkins methodology. He does an excellent job of illustrating the technique through real data examples. The best features of the book are the large number of examples for which Box-Jenkins time series models are fit. It also gives the beginner a nice undertanding of "how to do it" but not "why do it".
Rating:
Summary: nice text with many practical examples
Review: Pankratz gives a nice and non-rigorous description of Box-Jenkins methodology. He does an excellent job of illustrating the technique through real data examples. The best features of the book are the large number of examples for which Box-Jenkins time series models are fit. It also gives the beginner a nice undertanding of "how to do it" but not "why do it".
<< 1 >>
| 
