Synopses & Reviews
Statistical Learning from a Regression Perspective considers statistical learning applications when interest centers on the conditional distribution of the response variable, given a set of predictors, and when it is important to characterize how the predictors are related to the response. As a first approximation, this is can be seen as an extension of nonparametric regression. Among the statistical learning procedures examined are bagging, random forests, boosting, and support vector machines. Response variables may be quantitative or categorical. Real applications are emphasized, especially those with practical implications. One important theme is the need to explicitly take into account asymmetric costs in the fitting process. For example, in some situations false positives may be far less costly than false negatives. Another important theme is to not automatically cede modeling decisions to a fitting algorithm. In many settings, subject-matter knowledge should trump formal fitting criteria. Yet another important theme is to appreciate the limitation of one's data and not apply statistical learning procedures that require more than the data can provide. The material is written for graduate students in the social and life sciences and for researchers who want to apply statistical learning procedures to scientific and policy problems. Intuitive explanations and visual representations are prominent. All of the analyses included are done in R. Richard Berk is Distinguished Professor of Statistics Emeritus from the Department of Statistics at UCLA and currently a Professor at the University of Pennsylvania in the Department of Statistics and in the Department of Criminology. He is an elected fellow of the American Statistical Association and the American Association for the Advancement of Science and has served in a professional capacity with a number of organizations such as the Committee on Applied and Theoretical Statistics for the National Research Council and the Board of Directors of the Social Science Research Council. His research has ranged across a variety of applications in the social and natural sciences.
Review
From the reviews: "I believe that the practical utility of statistical learning over more traditional non- and parametric regression approaches has yet to be truly demonstrate but the procedures presented in this text do show considerable potential.... The mathematical prerequisites for using this book are minimal.... Some familiarity with using a computer is necessary in order to gain the most benefit for the text, and some previous experience of using a statistical software package would be advantageous." (C.M. O'Brien, International Statistical Review, 2009, 77, 1) "The readers of this book will obtain the knowledge of the dialectic of the regression modeling problems arising in the study of predictor -response. A large percent of the contents is devoted to discuss how to understand phenomena through regression equation fitting. ... I recommend it for practitioners and professors have the responsibility of teaching on the subject, the book gives an interesting perspective for dealing with regression." (Sovandep.H. Kumar, Revista Investigación Operacional, Vol. 30 (2), 2009) "On the positive side, SLRP is a nice addition to the data mining literature, more accessible than ESL. It gives good references and provides statistical detail. In general, I enjoyed the philosophical discussions about how statistical learning fits in with statistical inference. I may not hand it over to my colleagues in Biology and Sociology, but I will seriously consider recommending it to the undergraduates in my data mining seminar." (Richard D. DE VEAUX, The American Statistician, Novemeber 2009, Volume 63, Number 4, pp. 297-411) "...The strength of this book is its extensive discussion of practical issues. Algorithmic details are a starting point for discussing why and how methods work, comparison with other methodologies, limitations and strengths, and so on. Throughout the book, examples are worked through in detail. Each chapter except the first and the last end with a section headed 'Software Considerations', followed by 'Summary and Conclusions' and data analysis exercises. ...Regression methods, both the theory and the practice, remain a work in progress... .Berk has made a good start in pulling together commentary on issues of major importance." (Journal of Statistical Software, Vol. 29, Book Review 12, February 2009) "This book is unique in that statistical learning is discussed by a sociology-PhD scientist, Professor Richard Berk, who has extensive research
Synopsis
This book considers statistical learning applications when interest centers on the conditional distribution of the response variable, given a set of predictors, and when it is important to characterize how the predictors are related to the response.
Table of Contents
Statistical learning as a regression problem.- Regression splines and regression smoothers.- Classification and regression trees (CART).- Bagging.- Random forests.- Boosting.- Support vector machines.- Broader implications and a bit of craft lore.