Introductory Time Series with R (Use R!)

This is an excellent introduction to time series analysis in R, and is suitable for all readers who use R. In contrast to most statistics books, it does not presume an extensive mathematical background. Rather, it is a very much a progressive, didactic text, suitable for leisurely self-learning. The mathematics are presented briefly and appropriately for each topic, but progress and understanding do not depend on absorbing them in depth. It would be suitable, for instance, to social scientists, ecologists, public policy researchers, and so forth who use R. It is very much a multi-lesson tutorial on the basics of time series analysis, and should be worked through at the computer using R. The topics include decomposition (e.g., extracting seasonality vs. trends), handling autocorrelation, forecasting (e.g., the Bass model in marketing forecasts), regression models, and some more advanced topics such as spectral analysis. In some of the later topics, math is unavoidable and is presented when needed. There are two limitations to the book. First, as should be obvious from the preceding, some mathematicians and statisticians may be disappointed by the focus on tutorial rather than formal explanation. It has math but that's not the focus, so it would not be suitable for, say, a graduate-level mathematical stats course. Second, it of course cannot cover all aspects of time series analysis. It has examples from many domains (finance, operations, marketing, etc.) but limited depth in any single area; and it presents a variety of core models but does not cover the many advanced topics. Overall this is an excellent introduction to time series. If you're a general R analyst who wants to get started with time series, it's the best place to begin that I've seen.

Book is comprehensive but accessible to the non-statistician. Plenty of workable examples with simple code. Though not an R coding book, authors use pretty good coding practice and give some practical ideas for implementation. These guys clear up areas where I've previously struggled to get at the root of a method. In short, if you want to write TS proofs and academic papers, get another book. If you want to begin including more sophisticated TS models with your other work, this is the book for you. Only one knock, have been through several of the examples and there are some coding typos; nothing major but stay on your toes. Also, some of the algorithms may have changed since the book was written so you need to make use of the R help files to clear up any discrepancies with the author's work...some probably won't get cleared up because a more complex algorithm has been changed. This is a book I've been looking for.

A good book, although not presents a friendly and logical order to the subject. It has a good data set to students work throughout book. A summary of R code is shown at the end of chapters. Um bom livro para começar estudar séries temporais. Apesar de não ter uma sequência bem definida nos assuntos abordados, encontramos muito conteúdo no livro. Recomendável, principalmente por ser bem fácil obter o R a partir da internet, o que facilita o aprendizado quando não se dispõe de softwares consagrados, mas muito caros.

This is a cracking book on applying R to time series analysis. The best parts of the book are all of the worked examples, the accompanying data sets and several different ways to calculate seasonality. The book is better than most on time series, because it does not neglect the de-trending process needed to get stationery residuals. If you use just the lm() command in R to do this before, then the real gem in this book is the advice to use the gls() command from the nlme library instead (to get the confidence intervals right). Overall, a very good book that is applied to R but has enough mathematical backing for the techniques presented. However, this is a book about applying time series analysis in R. If you seek a more algebraic treatment, then this is not the book I'm afraid, but it would be a great supplement!

wow what a poorly written book. First it does not specify how to obtain the functions.In order to use functions you need to know what is the "package" the function is in, if you do not know the package then what, google it?. Second, terrible mediocre statistical explanation. If you are not familiar with stats you won't understand anything. I am sending it back right now.

I used this in a graduate level course and found the structure of the subjects useful and the addition of R syntax marginally synthetic, particularly in appendices. However, many of the topics covered in chapter exercises include tasks in following chapters; this is NOT useful! Additionally, some of the syntax used in Exhibits are inconclusive and only marginally helpful e.g. incomplete script [clearly a editorial oversight]. When prescribing to this text be prepared to use Cran's R directories references for assistance; R proficiency dependent. This is a sound text likely as supplemental material but not exclusively.

The book assumes you have some hands-on experiences with time series modeling. Otherwise, you may get lost in some topics. It is a high-level short review of most common topics and models in the classic time series analysis. For example, the book didn't talk about Recurrent Neural Networks (RNNs) or long short time memory (LSTM) cells that are very suitable for some classes of time series projecting/classification. I didn't want to reproduce the R code examples in the book and just read the codes as regular text and it was pretty informative. I eventually used Python libraries for my project. I highly recommend this book.

Great comprehensive introduction. Dont let the word introduction fool you. Sometimes it fools me into believing a book is going to be tok basic and a bore. This book avoids that trap and covers a heap of content whilst striking a balance between technical and intro level understanding. I actually look forward to time alone so i can read it.

This is a most excellent text. The authors combines meticulous pedagogical care with the practitioners' experience.

As stated in the title, this book introduces the reader to a number of time-series analysis methods supported in R. Several packages are presented, and the time-series format. There is a lack of coverage of the most advanced, and therefore more commonly used time-series formats, such as zoo, but I expect that the novice programmer will still find in this book a lot of useful information (and examples) on the various time-series methods and their application through R.

Mit R ist wohl eine Revolution in der angewandten Statistik und Modellierung eingeleitet worden. Analytic Tractability ist nicht mehr das Kriterium. Man kann es auch mit dem Computer simulieren. Dieses Buch ist ein sehr gutes Beispiel für diese Revolution. Die Autoren machen in 12 Kapiteln und 250 Seiten einen relativ umfassenden Streifzug durch die verschiedenen Time-Series Methoden und die dazugehörigen R Funktionen. Es stehen nicht wie früher abstrakte mathematische Sätze sondern konkrete Modelle und Resultate im Vordergrund. Man kann sich selbst mit den Zeitreihen spielen und so durch unmittelbare praktische Übung ein Gefühl für die Methoden bekommen. Die handwerkliche Seite der Statistik bekommt den ihr zustehenden Rang und Würde. Man könnte über jedes Kapitel ein eigenes Buch schreiben. Die Autoren konzentrieren sich aber auf das Wesentliche. Das sonst so lästige und das gibts auch noch, und das auch, for further details see ... fällt weg. Dadurch hat man am Ende einen doch sehr umfangreichen Werkzeugkasten mit dem man Einiges anstellen kann. Ich kannte die Methoden von der Theorie her (siehe Literaturliste) und habe auch schon Einiges per Hand in C implementiert. Mit einem derartigen Hintergrund ist das Buch ein Genuss zu lesen. Und selbst als alter Hacker schwört man sich, daß man in Hinkunft R und nicht mehr seine selbstgebastelten C-Routinen verwendet. Wahrscheinlich ist das Buch aber für einen absoluten Anfänger doch etwas zu kompakt. Ein sanfter Einstieg ist: P.J.Brockwell, R.A.Davis: Introduction to Time Series and Forecasting. Leider verwendet das Buch die proprietäre und umständliche ITSM2000 Software. Ein Klassiker ist: G.P.Box, G.M.Jenkins, G.C. Reinsel: Time Series Analysis. Interessant fand ich auch: DSG Pollock: A Handbook of Time-Series Analysis, Signal Processing and Dynamics. P.S.: Eine kleine Ironie der Geschichte ist, dass ausgerechnet die Wirtschaftsuni Wien bei der R!evolution eine wichtige Rolle spielt. Mit Prof. Hornik gibt es einen Gulliver unter den Betriebswichteln der massgeblich an der Entwicklung von R beteiligt ist.

Ich habe das Buch im Kindle und habe festgestellt, dass die Datensätze jetzt an einem anderen Ort zu finden sind, da der Autor wohl die Uni gewechselt hat. Da Amazon Links aus den Rezensionen heraus kürzt, hier also eine "Anweisung", wie man die Daten finden kann: Statt die ursprüngliche URL anzugeben, die sich an der Uni Massey (massey.ac.nz) befindet, die URL dahingehend abändern, dass man jetzt direkt nach dem doppelslash die neue Uni (ohne www) dort eingibt: (elena.aut.ac.nz) und den Rest der URL konstant lässt. Beziehungsweise die neue Homepage von Hrn Cowpertwait suchen und dann noch Slash und ts an die URL der Homepage anfügen.

The contents of the book is well chosen, but unfortunately the book cannot be recommended, because nearly none of the datasets is available anymore. Therefore it is not possible to do most of the exercises.