<< 1 >>
Rating: 
Summary: Brilliant exposition and crystal clarity
Review: This book is a breath of fresh air for anyone who is attempting teach time series analysis to themselves; especially the Spectral Analysis techniques. The book has three primary strengths. First, it assembles between two covers EVERYTHING that one needs to understand the developement of Spectral Methods. Secondly, the authors write in a very clear manner with no attempt to obfuscate the material. The authors clearly want to teach these techniques rather than satisfy some publish or perish requirements. Thirdly, and perhaps most importantly, the book has detailed and completely worked out examples using actual numbers in the matrices. One may think symbology should be enough but actual numbers make it very easy to trace the logic . This is a terrific tome and destined to becoming a teaching classic. It has generrated the longest series (no pun intended) of"Ah Ha!" experiences for me that I have had the pleasure of in a very long time. Why can not every text explain singular value decomposition the way these authors have? Showing the link between difference equations and differential equations and the segue into them is a work of art. I am very pleased with my purchase
Rating: 
Summary: Useful (although expensive) introduction to SSA
Review: This text is divided into three main parts: 1) Mathematical notes, 2) Theory and Methods, 3) Applications. As a reader interested in time series analysis but unfamiliar with singular spectrum analysis (SSA) per se, I was not really sure what SSA was exactly, or what it is suppose to accomplish, until nearly halfway into the book. Very early on, I found no one single sentence, paragraph, definition, or description that clearly defined SSA beyond a "technique based on spectral decomposition."Although the text is clearly written and well organized, the authors' target audience is not well established. For example, Section I starts with an rather remedial primer on linear matrix algebra, yet Section II already seems to assume the reader has a fairly mature knowledge of statistical inference, hypothesis testing, and principal component analysis (which might be gained from courses on statistical analysis of variance, or ANOVA). The authors also make frequent references to other time series analysis techniques, such as ARMA modeling and Fourier analysis in Section III. Familiarity with ANOVA and other kinds of time series analyses will certainly benefit the reader. One gets the impression that the manuscript has been enhanced to fill a textbook-sized volume (and would explain why the book dedicates an entire chapter to numerical examples illustrating such things as how to multiply two matrices). The font size is strikingly large for a 6" x 9" textbook, and the layout is noticeably fragmented with lots of section subtitles and large amounts of white space around these section titles. Specifically, each page accommodates no more than 34 lines of 4 1/4" wide text, which I compared with several textbooks this size which usually accommodated 40-45 lines of 4 1/2" wide text (I suspect the 153 page content might have been fitted into about 100 pages just by changing the book layout). Given the quite basic subject matter of the first few chapters, the remainder of the book is but a fairly short survey on the main subject of SSA. The advantage is that its contents can be covered relatively quickly. The rarity of introductory texts specifically dedicated to this niche subject matter, and the uncomplicated presentation the authors have chosen, make this book a worthwhile, albeit expensive, introduction to the topic. This book has nice presentation qualities, and had it been issued as a thin, inexpensive paperback, this would more easily be a 4 or 5 star book. Similarly, this is a 4 star book for the prospective reader analyzing data related to the atmospheric sciences or climatology (as much of the authors' background materials emphasize these areas). An expanded second edition would be welcomed.
<< 1 >>
| 
