Data Mining and Business Analytics with R

This is a timely and excellent book. Its greatest strength lies in the carefully presented statistical models coupled with diverse and interesting real-world examples. Ledolter effectively sets the stage in Chapter 1 for what is to follow by explaining the difference between traditional statistics applications and the problems for which data mining techniques are necessary. He highlights the nature of data mining problems and describes the techniques for addressing them that are discussed in subsequent chapters. The early chapters review traditional regression and logistic regression models with applications. Then the book moves quickly to lesser known techniques that are particularly useful for dealing with large data sets. These methods include nearest neighbor analysis, Bayesian analysis, regression and classification trees, clustering, and market basket analysis. The book ends with a comprehensive set of exercises. The last eight of the exercises are particularly valuable because they provide detailed worked examples and in a number of cases include alternative statistical approaches to the same problem. All of the final exercises are tied to the book's chapters, while all examples and exercises make use of the powerful and free R Statistical Software. The complete R code is available on the book and author websites.

Back in school and going after a second Master's degree now. Needed this book for the class material. I find the class enjoyable enough, and this book is definitely helping me understand the class concepts.

Book in Great Condition, Arrived Early

When I saw the title "Business Analytics" I thought that this might be a book that targeted MBA students who are uncomfortable with graduate level applied math and statistics. Books that follow this pattern provide a cook book approach to packages that do linear regression or clustering, without providing much background. The virtue of these books is that they tend to be more readable than book like The Elements of Statistical Learning by Hastie et al. Elements of Statistical Learning is a classic but it covers complex topics in a few paragraphs or a page. There are parts of this book that I have read again and again before fully understanding the material. Data Mining with Business Analytics is a much gentler introduction to many of the topics in "Elements" (statistical analysis, linear models and clustering, among other topics). Johannes Ledolter writes clearly and walks the fine line of discussing the mathematical background without providing a deep discussion of the mathematics. As the title suggests, the examples are in the R statistical language. I have been using R for several years and have become an R fanboy. I see R as an indispensable platform for doing data analysis. The R examples are generally well developed. R includes a number of data sets and many authors use these data sets to illustrate analytic techniques. I am starting to feel that the prostate cancer data set, which is used in this book, is getting a bit old and I propose a moratorium on its use. When I studied linear modeling we covered a lot of the mathematical formalism and proofs that this book leaves out. While this did give me a deeper understanding of linear modeling, the cost was some topics, like logistic regression were omitted. This book has a good chapter on logistic regression. Logistic regression is probably the most popular way to analytically do credit analysis. The logistic regression chapter includes examples of logistic regression applied to lending and credit. The book does not "talk down" to the reader. A basic background in statistics is assumed. One of my professors said once that "MBA students don't like linear algebra" and I found it interesting that topics like linear regression were presented without linear algebra (e.g., as finite math using summations). The book provides a solid introduction on the techniques and their implementation in R. For anyone using these techniques this will serve as a starting point. For example, the discussion of K-nearest neighbors clustering gives the reader a feel for clustering. For many applications, clustering is more complicated and there are books on this topic. The prices of math and programming books can be hard to swallow. I think that most readers who want an a solid overview of data mining will find that this books does pay back its substantial cost.

This graduate-level textbook gives students very good exposure to the use of open-source statistical software R in data analysis, data exploration, and data model construction. Readers must already know some R basics (e.g., how to install R packages, read help files for packages and functions, and work with basic R data structures such as data frames, etc.) and statistical concepts such as hypothesis testing, significance levels, etc. The book chapters are organized mostly around statistical techniques such as linear regression, clustering, text analysis and social network analysis. Each chapter usually begins with a discussion of concepts important to understanding the statistical technique in question, followed by descriptions of the datasets and R packages to be used in the hands-on problem-solving exercises. By following along, readers will acquire useful knowledge on what data modeling problem(s) a particular statistical technique can be applied to, what pitfalls (e.g., overfitting) to avoid, how to utilize the covered R packages and use the provided code examples as templates for studying similar problems. The book has an "applied" emphasis -- discussions of the mathematical details underpinning a statistical technique are kept to a minimum and to a relatively high, conceptual, and practical level. The datasets are quite varied, covering a wide range of domains relevant to the fields of engineering, business and marketing, economics, and health care. All datasets and code examples are available for download from websites mentioned in the book. The code examples have comments, but there is room for improvement (for example, readers who are relative R novices may not know why a call to set.seed(x) is required before calling some specific R functions). A similar observation can be made regarding the graphics presented in the book: providing figure captions and, in some cases, better x- and y-axis labeling (for example, instead of just labeling the x-axis with a 0 and a 1, use labels that indicate what the 0 and the 1 represent or mean), in my opinion, could help enhance the reading experience. Overall, however, I thought the book is written at a level that its intended audience will find accessible.

The book is good, but you really have to download the code. The author skips big blocks of code in the written text, so you can not follow along, entering R commands, by reading the book in isolation. The author in many cases offers sparse explanation for the technique and the analysis he offers is quite curt as well. Overall, it is a good book if you know stats, know R, and have a lot of time to study the book in conjunction with his notes. It is extremely tedious, but then again, consider the subject matter. Machine Learning by Lantz suffers from none of the issues I point out here and I would recommend starting with that first and then follow on to this.

While this book is expensive, if you know some R already and you are looking for more examples for covering statistical learning/modeling methods, with sample code, this is an excellent buy. If you don't know R you will need to spend a bunch of time getting up to speed before you hit this book. Even though it is targeted at the business world the examples and code are widely applicable. From the table of contents and index you can get a feel for the topics covered. If you want to see the actual code used check the books website. It is solid and includes the code, the data sets and an errata. If you are curious to know what R libraries are touched, they include: arules, car, class, cluster, elipse, igraph, lars, lattice, leaps, locfit, MASS, mixOmics, mixtools, nutsheell, ROCR, startnet, textir, tree and VGAM. This is a pretty book with great well annotated graphics to help you learn. The writing is pleasantly clear and direct without being too terse. While the code could be better commented (for the R novices) in general it is good and the text which surrounds the code is very good. There are formulas here. The math complements the writing rather than being a deep dive. The references to outside work are on target but the author does not include some obvious choices like An Introduction to Statistical Learning: with Applications in R (Springer Texts in Statistics) for a deeper look at the math or Data Analysis and Graphics Using R: An Example-Based Approach (Cambridge Series in Statistical and Probabilistic Mathematics) for additional examples. Overall, this is an excellent book for someone who knows basic statistics and the fundamentals of R and who wants to learn modern methods using examples.

This is meant to be a practical book. The author's "objective is to provide a thorough discussion of the most useful data-mining tools that goes beyond the typical 'black box' description, and to show why these tools work". I think the result of reading and doing the exercises in this book is: 1. I will have acquired some familiarity with regression techniques and a few of the problems they can help with 2. I will have performed the regression techniques in R Over half the text focuses on various kinds of regression. Then there is a little bit on classification, decision trees, clustering, principal components analysis. However, I also think: 1. The math is so fast --- mostly one definition after another --- that its inclusion is superfluous. If you know what the equations are, there is no need to see them here. If you don't, this presentation isn't a good way to learn them. These sections often say to check out the author's 2006 text Introduction to Regression Modeling for more details. 2. The mentions of alternate software (Minitab, SAS, SPSS) are useless throwaways and should either be removed or expanded. Who cares if I have two brands of calculators that give me the same answer for 3 + 4? Likewise, there is no need to say that R gives the same answer as Minitab in a single example (pp. 88-92) or that some feature exists in SAS and SPSS (CHAID, p. 186). 3. The exercises are extremely important for practicing. 4. Examples sometimes have long program output. In my experience, it takes some practice to read the program output and understand what each number means, and this discussion is not really done in the text. From a statistics perspective, I would instead recommend Tibshirani, Hastie, Friedman: The Elements of Statistical Learning . For machine learning techniques, Segaran: Programming Collective Intelligence .

This is a book with a lot of content and so I haven't read it all. Instead I've sampled some chapters that looked interesting. This is what I've noticed. It's in colour! This is great for showing the graphs that can be done with R, with colour used. It's the only book of its kind I can think of that has that. The writing is clear. The author seems to know what I am thinking and supplies the right information at the right time, mostly. It's much better than most mathematical/statistical texts. There are lots and lots of detailed examples in R that you can load up and try out. The topics develop from one chapter to the next, but it doesn't quite have the typical text book coverage. Instead, the methods chosen are more varied and seem to me to reflect the practice of data mining in reality. Those are all good points and I think this book deserves five stars.

Difficile trovare un libro sull'argomento del data mining e analytics che sappia conciliare la teoria con la pratica. Questo libro rappresenta un buon esempio di quanto, esponendo in modo chiaro ed efficace la teoria e, a seguire, esempi di applicazione con codice in linguaggio R. Peccato che l'autore non abbia avuto modo di esporre ogni metodo con la stessa valida efficacia, per cui la qualità di esposizione di alcuni metodi non e' allo stesso livello di profondità di altri (es. regressione vs. SVM). Libro valido anche per la quantità di metodi di data mining esposti, anche se alcuni argomenti meriterebbero ancora maggiore dettaglio.

This book aims to be graduate-level, and in terms of the explanations given for the various concepts the level of statistical understanding assumed is consistent with that. Having said that, someone with less of a solid mathematical background could potentially read through the overview of various concepts and follow the examples to get a working knowledge of how to create a range of plots of data and perform classification, dimension reduction, network analysis and more. To be clear on what it doesn't cover - this is not an introductory text for R - the examples assume a working knowledge of R, so without this they may be hard to follow. Sophisticated terms are regularly used without explanation, so you may also (like myself) find yourself having to look up the occasional term. It also doesn't try to cover things like bias as data collection in general is outside the scope of the book - it assumes suitable data for analysis already exists. The book is concise (to the point of losing context and clarity on occasion), meaning that despite being only a few hundred pages or so, it covers a significant range of techniques, both in outlining the theory and mathematical basis and in worked examples. These examples show graphical and text outputs and account for possibly half of the book, but despite this it is hard to say that it's overdoing this because it keeps the scripts as brief as possible while illustrating practical use, and this includes loading available datasets, performing analysis and showing useful outputs, whether graphical or text. Recommended if you're up for a challenge and feel you can follow sophisticated concepts.

I have been a user of the R statistics program for a number of years. One of the problems of R is that it isn't very easy to learn and it does not have an effective user interface like SPSS or Excel and so getting data in has been the problem. Once you have data there then R has a huge wealth of statistical analysis techniques that you can use. The challenge then is to find the right one for you. This is where this book is invaluable in showing you how to use R for business analytics. It gives you all of the methods and the techniques that you should apply to your data. Hopefully this will encourage more users from a wider range of subjects to use R. By giving "recipes" for data analysis this book saves you the time and difficulty of going through the somewhat impenetrable online documentation. By using an IDE like R-studio it is also becoming much easier to manage the packages and the command line interface. So I think that this is an invaluable resource for anyone who wants to use R for their analytics, but it is not a general text on analytics themselves.

My main gripe/concern with this book is the title. In particular one word. The word 'with'. See, the problem is that no matter which way I look at it, 'with' should say 'in'. This isn't a theoretically bound volume which you can trawl through whilst wishing you had a higher IQ. However, it does serve as a good introduction to using R in order to evaluate various sets of data - if you are new to Data Mining, take this book and read it alongside another (i.e. Witten and Frank's 'Data Mining' book that also has an emphasis on Java). Otherwise, if reviewing this as an introduction to the facilities that R has to offer for Data Mining and Analytics, its top notch. I especially like the chapter on network data - lots of things can be done in R that correspond to Gephi and can be quite well compared.