Applied Predictive Modeling by Max Kuhn - Book Review


The overall predictive modeling method is covered by Applied Predictive Modeling, starting with the critical steps of data preprocessing, data splitting, and model tuning foundations. The text then offers intuitive examples of various traditional and modern techniques of regression and classification, always focusing on explaining and solving specific data issues. Addressing practical issues goes beyond model fitting to topics such as handling class imbalance, choosing predictors, and recognizing causes of the model's poor performance, all of which in practice are commonly occurring issues.

Via several hands-on, real-life examples, the text explains all aspects of the modeling process. And each section includes a detailed R code for each phase of the method. The data sets and the required code are included in the AppliedPredictiveModeling R kit of the book's companion, which is freely available on the CRAN archive.

This multi-purpose text can be used to introduce predictive models and the overall modeling process, a practitioner's reference manual, or a predictive modeling text or courses at advanced undergraduate or graduate levels. Each chapter includes problem sets to help solidify the topics covered and uses information available in the book's R package.

Any fundamental knowledge of R. should be accessible to readers and students interested in applying the methods. And some mathematical knowledge is needed for a handful of the more advanced topics.


At RStudio, Max Kuhn is a software engineer. He is focusing on enhancing the modeling capabilities of R at the moment. At Pfizer Global R&D in Connecticut, he was the Director of Nonclinical Statistics. For over 18 years, he has been developing models in the pharmaceutical and diagnostic markets. Max earned a Ph.D. in Biostatistics. Max is the author of several R packages for machine learning and reproducible analysis techniques and is an Associate Editor for the Statistical Software Journal. He and Kjell Johnson wrote the Applied Predictive Modeling book, which received the American Statistical Association's Ziegel prize, which honors the best book reviewed in 2015 in Technometrics. In 2019, their most recent book, Feature Engineering, and Selection was released.

Dr. Max Kuhn is Director of Non-Clinical Statistics for Groton, Connecticut, at Pfizer Global R&D. He previously worked on producing molecular diagnostics for infectious diseases at Becton Dickinson (BD). For predictive modeling, Max is the creator or maintainer of many R packages: Caret, AppliedPredictiveModeling, SparseLDA, Cubist, C50. He teaches predictive modeling classes regularly at Predictive Analytics World and Person! Work on biomarkers in neuroscience, drug discovery, molecular diagnostics, and response surface methods are included in his publications. He can be contacted at [email protected] via email.

Dr. Johnson has over a decade of work in statistical analysis and predictive analytics in pharmaceutical research and development. He is a co-founder of Arbor Analytics, a predictive modeling company, and is a former Director of Statistics for Global R&D at Pfizer.  His academic work focuses on applying mathematical methods and studying algorithms, and improving them.

Buy Now!

Post a Comment